Recent work in acoustic landmine detection has shown that many landmines exhibit a multi-mode vibration pattern. To fully map the vibration pattern of these modes requires spatial resolutions on the order of millimeters. An optical technique that lends itself to such vibration sensing is an electronic speckle pattern interferometer (ESPI). In this work the double-pulse ESPI system has been used for the vibration measurement of the ground surface. The principle of method is based on recording two specklegrams of the object with two laser pulses synchronized with the vibration peak and the vibration valley respectively. The 2D vibration amplitude spatial distribution is obtained by subtracting two specklegrams and processing the received correlation fringe pattern. The experimental setup uses a mechanical shaker to excite vibrations in the ground to significantly increase the vibration amplitudes at the spot of interest and a laser Doppler vibrometer to detect the resonant frequency of the mine. Experimental results are presented from laboratory experiments. The spatial maps of the vibrating ground over buried antitank and antipersonnel landmines are studied. The effect of the vibration of a granular material like sand on the speckle decorrelation is discussed. This material is based upon work supported by the U. S. Army Communications-Electronics Command Night Vision and Electronic Sensors Directorate under Contract DAAB15-02-C-0024.

Resonance behavior of many types of landmines was first experimentally discovered in 2000 (Donskoy et al. in Proceedings of SPIE Vol. 4394, pp. 575-582, 2001). Laboratory studies and field tests have shown that mine’s resonance response is a complex phenomenon dependent upon interaction between soil and mines and their respective properties. Although the resonance effect was successfully used by various research teams for detection of landmines, there were no thorough studies on various factors influencing buried mine's resonance response. This paper presents results of theoretical and experimental investigation of this problem including multi-modal structure of mine's vibration response, effect of burial depth and soil condition. In the modeling efforts we considered multiple modes of vibration of mine casing and represented them as oscillators with effective parameters. This approach allowed for simplification of analysis and expanding existing lump-element model to account for multiple vibration modes. The experimental tests were focused on studying the effects of burial depth and soil moisture content on resonance behavior of soil-mine system. The tests have shown that a resonance frequency initially decreases with burial depth, as expected. However, an anomalous resonance frequency increase was observed at greater depths; soil moisture even further increases the resonance frequency.

The paper presents measurements taken with a scanning ultrasonic Doppler vibrometer on a landmine buried separately in sand and in grass-covered soil. The signal obtained with a laser Doppler vibrometer experiences a large variability that is due to loss of spatial coherence upon scattering from moving grass blades. Ultrasonic sensing is not affected by this limitation since the acoustic speckle is much larger than its optical counterpart. Moreover, the slightest hint of air motion enhances the motion of the grass blades, which adds to the optical decoherence and subsequent loss of useful signal. It is shown also that the ultrasonic system has no problem penetrating the layer of grass and detecting the location of the buried target excited by a mechanical shaker.

The coupling of airborne sound into roadways and desert soils has been significantly investigated for the purposes of locating buried antitank (AT) landmines. However, there has been relatively little acoustic-to-seismic (A/S) coupling data collected for the purpose of buried antipersonnel (AP) landmine detection. A/S coupling landmine research has typically been accomplished with a low frequency sound source radiating pseudo-random noise in the frequency range 0f 80-300 Hz and a scanning single beam laser Doppler vibrometer (LDV) with a 10 cm beam spacing. The single beam LDV is operated in a serial data collection procedure resulting in long scan times. Recently, a data collection platform that uses 16 LDVs has been used to detect buried AT landmines. In the work reported here, this data collection platform is used to scan a significant number of AP landmines. For this purpose, the LDV beam spacing is reduced to 3 cm and the upper frequency of the sound source is increased to 2 KHz.

In nonlinear acoustic detection experiments involving a buried inert VS 2.2 anti-tank landmine, airborne sound at two closely spaced primary frequencies f1 and f2 couple into the ground and interact nonlinearly with the soil-top pressure plate interface. Scattering generates soil vibration at the surface at the combination frequencies | m f₁ +- n f₂ | , where m and n are integers. The normal component of the particle velocity at the soil surface has been measured with a laser Doppler velocimeter (LDV) and with a geophone by Sabatier et. al. [SPIE Proceedings Vol. 4742, (695-700), 2002; Vol. 5089, (476-486), 2003] at the gravel lane test site. Spatial profiles of the particle velocity measured for both primary components and for various combination frequencies indicate that the modal structure of the mine is playing an important role. Here, an experimental modal analysis is performed on a VS 1.6 inert anti-tank mine that is resting on sand but is not buried. Five top-plate mode shapes are described. The mine is then buried in dry finely sifted natural loess soil and excited at f₁ = 120 Hz and f₂ = 130 Hz. Spatial profiles at the primary components and the nonlinearly generated f₁ - (f₂ - f₁) component are characterized by a single peak. For the 2f₁+f₂ and 2f₂ + f₁ components, the doubly peaked profiles can be attributed to the familiar mode shape of a timpani drum (that is shifted lower in frequency due to soil mass loading). Other nonlinear profiles appear to be due to a mixture of modes. This material is based upon work supported by the U. S. Army RDECOM CERDEC Night Vision and Electronic Sensors Directorate under Contract DAAB15-02-C-0024.

Acoustic landmine detection (ALD) is a technique for the detection of buried landmines including non-metal mines. Since it gives complementary results with GPR or metal detection, sensor fusion of these techniques with acoustic detection would give promising results. Two methods are used for the acoustic excitation of the soil: laser excitation and loudspeaker excitation. A promising concept is using lasers for excitation and monitoring for complete stand-off detection. Results from a field test and laboratory experiments show the feasibility of laser excitation for ALD. In these experiments buried landmine surrogates were measured with ALD using a Nd-YAG laser at 1.06 μm for the acoustic generation and a Laser Doppler Vibrometer (LDV) system at 1.54 μm for the detection of soil vibrations. An analysis is given of the experimental results showing the potential and the inherent limitations of the technique. We discuss the relative merits of LDV detection versus microphone detection of the laser-induced acoustic vibration. It was found that the LDV has limitations with respect to microphone detection due to the influence of surface effects that are prominent in LDV but absent in microphone detection.

Acoustic-to-seismic coupling technology using an LDV as a vibration sensor has proved itself as a potential confirmatory sensor for buried landmine detection. One of the most important objectives of this technology is to increase the speed of measurements over traditional point-by-point scanning LDVs. A moving cart that uses 16 LDVs as well as a continuously-scanning single beam LDV have recently been demonstrated to increase the speed of detection. Recently a multi-beam LDV simultaneously probing 16 positions on the ground has been developed and successfully used for landmine detection. In this work, we report on a continuously-scanning multi-beam LDV as a confirmatory sensor for acoustic landmine detection. The multi-beam LDV simultaneously illuminates the ground in 16 points spread over a 1 meter line. A scanning mirror moves all 16 laser beams across the line. The system enables scanning a 1 meter square area in a much shorter time than previous scanning techniques. This material is based upon work supported by the U. S. Army Communications-Electronics Command Night Vision and Electronic Sensors Directorate under Contract DAAB15-02-C-0024.

The utility of acoustic-to-seismic coupling systems for landmine detection has been clearly established. In this approach, laser Doppler vibrometers (LDV) are used to measure the different responses to acoustic excitation in ground regions with and without buried landmines. Currently, for most applications, only the magnitude of the surface velocity is investigated and used to construct recognition algorithms. Recently, we introduced phase-based features in the classification scheme, significantly lowering false alarm rates at given detection probabilities. In this paper, we present modeling equations that explain the phase features for ground areas both with and without buried landmines from the perspective of harmonic oscillator models. We also describe the image processing techniques applied to velocity data collected in the time domain with a moving LDV array. The observed signatures are also compared with the prediction of the models described. We also construct classifiers with only magnitude information and both magnitude and phase information for this time-domain data set. Classification results indicate that we can combine magnitude and phase features to improve the detection of buried mines while reducing false alarms. We also find that using phase information improves the distinction between ground regions with buried landmines or man-made clutter objects.

The combination of a powerful acoustic transmitter with high resolution laser spectroscopy has led to a promising approach for the detection of buried landmines, especially for those with no or only minor metal content. This paper summarizes current R&D work on this new technology including a brief sensor overview and a more detailed description of our data processing methods. The SLDV sensor picks up tiny vibrations of the soil surface in the order of some μm/s in a rectangular grid of measuring points. We use a multi-threshold algorithm for the segmentation of mine cues and reduce false alarms by analyzing the stability of object size, contrast and shape in the frequency domain. In addition to the amplitude of soil vibration the phase is investigated as a secondary information channel to optimize the classification procedure.

Exciting the ground with an acoustic tonal projected by a loud speaker is one method for detecting buried landmines. The subsequent ground motion is measured with a laser Doppler vibrometer (LDV). The LDV data contain the tonal in a frequency modulated form. One approach for demodulating the data and extracting the tonal uses a Hilbert transform. The ground velocity can be obtained from these data to identify mine presence or absence. An alternate approach to mine detection is to perform consecutive fast Fourier transforms on the modulated LDV data, and to average the output powers in each spectral bin. This results in a ground velocity distribution function in the spectrum that is manifested by a broadband of modulated frequencies. The proximity of the beams to a mine (over, near, not near) can be determined from the bandwidth of the modulation. Furthermore, the velocity distribution functions provide additional information that previous techniques do not. Such information may be useful for separating mines from false targets. This technique is discussed, and the results from measured MB-LDV data are presented. This paper is based upon work supported by the U. S. Army Communications-Electronics Command Night Vision and Electronic Sensors Directorate under Contract DAAB15-02-C-0024.

This paper addresses a potential method to advance acoustic landmine detection by increasing operator and equipment standoff range from the minefield and developing a lightweight system that is potentially more practical than many currently researched systems. In this study, a parametric array acoustic source is evaluated to understand its potential for landmine detection. The array can transmit audible signals over 100 meters in air and has a weight of just four pounds. A proof-of-concept system was built at M.I.T. Lincoln Laboratory that uses a commercial parametric source to insonify the ground and excite buried mines. A commercial laser vibrometer was then used to measure the displacement velocity at the ground surface on and off the mine. This system has been demonstrated at an outdoor landmine facility and has measured signatures from buried anti-personnel mines. The overall concept shows promise; however, the parametric source used in this preliminary test was developed for home entertainment and will require substantial modification to be practical for landmine detection.

Inversion methods for estimation of geoacoustic model parameters often use the scattered field data for obtaining the properties of viscoelastic layered media. This work presents a method to retrieve soil background parameters using the outdoor acoustic-seismic transfer function (admittance). Clutter in landmine detection is related to with spatial variations of soil parameters, so knowledge of soil parameters and their spatial variability are very important for landmine detection. The resonance method is extended and used for preliminary estimation of a set of parameters for a three-layered ground model. The least squares method is later used to choose the model with the best fit to experimental data. Results of the reconstruction show good agreement with the experimental data. A description of the resonant technique and the experimental setup are presented. The effect of a finite size of the sound sources often used in acoustic landmine detection on the acoustic-seismic transfer function is also discussed.

Experimental and numerical models have been utilized at Georgia Tech in the research and development of a seismic landmine detection technique which generates seismic waves in the soil using a surface-coupled electrodynamic transducer and detects normal surface displacements with a non-contact radar sensor. As the numerical models have shown a strong dependence upon material properties of the soil as a function of depth, experiments have been conducted at six field sites and in the experimental model to quantify the effect of different soil conditions upon the operation of the seismic landmine detection system and to measure depth-dependent material properties. Measurements have been made with and without buried anti-personnel and anti-tank mines to determine the effects that landmines have upon the propagation of seismic waves. Surface waves have been measured using the non-contact radar sensor as well as triaxial accelerometers and geophones. Post-processing has included the examination of particle motion in three dimensions, the identification of individual wave types through polarity tracking and dispersion curves, and the extraction of individual propagating waves. The field sites include wet and dry sand at a beach, a roadbed at a U. S. Government facility in a temperate climate, frozen ground, clayey soil with and without rocks, and a silt-sand mixture in a coastal region.

This paper addresses the design of time-domain signals for use as seismic excitations in a system that images buried landmines. The goal of the design is the selection of a signal that provides sufficient contrast for the post-processed landmine image in the shortest possible measurement time. Although the goal is relatively straightforward and the problem appears similar to one of system identification for a linear time invariant (LTI) system, practical implementation of many commonly accepted approaches to the system-identification problem has proven difficult. The reason for this is that the system under consideration exhibits observable nonlinearity over the entire range of drive levels that are of interest. The problem is therefore constrained by the requirement that nonlinear effects be tolerable rather than imperceptible (i.e. that the nonlinearity be sufficiently weak that the system can be reasonably characterized as linear). Several candidate signal types that have been shown to offer good noise immunity for the LTI system identification problem were considered. These included circular chirps, binary-sequence-based (BSB) signals, and numerically optimized randomly seeded multisines. Based on purely experimental figures of merit, circular chirps with flat amplitude and linearly swept frequency offered the best performance among the signals that were tested.

Impulsive vibration of the ground can generate seismic interface waves of the Rayleigh type, which decay exponentially with depth into the soil, and spread cylindrically with lateral range. At useful frequencies around 100 Hz, they typically travel at speeds around 100 m/sec, with wavelengths around a meter. Rayleigh waves can be made to propagate in sonar-like pulses to buried targets, reflect, and return to the sonar for reception and signal processing, providing range, bearing, and information as to target type. We have conducted new experiments and analyses with seismic sonar in a clay soil. A focused array of 10 sources and 8 receivers (tri-axial seismometers) were deployed at a range of 4.5 m to examine a 20 lb. landmine as well as a clump of rocks, and other false targets. After vector polarization processing, the amplitude of the mine target echo was 28 dB above the environmental backscatter. Mine-like target confirmation was provided by cross-Wigner-Ville transformation applied to polarized echoes. A first attempt to discriminate man-made from natural targets by identifying energy patterns on this evolutional cross-spectrum approach is presented. The potential for target detection as well as a level of target type classification, at relatively long ranges, was demonstrated.

Experimental measurements have shown that the use of a multi-layered elastic media is necessary for transfer function numerical modeling. The present work deals with the effect of variability of ground properties (compression and shear wave speeds, density, attenuation and thickness of the layers) on the acoustic-seismic transfer function (admittance) and on clutter in landmine detection. Analysis is performed on the planes of parameters of the ground in a wide frequency range for all angles of incidence. Matrix approach is used to increase the accuracy of computations. It is revealed that the acoustic-seismic transfer function is sensitive to ground properties and that small variations in the shear speed may cause strong variation in the acoustic-seismic transfer function. Results of outdoor measurements of the acoustic-seismic transfer function are presented and a correlation between high magnitudes of the acoustic-seismic transfer function in certain frequency ranges (false alarms) and moisture content on the surface is revealed. A simple model explaining the correlation between moisture content in the upper layer, acoustic-seismic transfer function and ground properties is suggested.

Time Reversal is based on the fact that most physical laws of nature are invariant for time reversal, i.e., when time t is replaced by -t, most physical laws remain unchanged. Physically this means that by time reversing, a particle will retrace its original path or trajectory. Based on this fact, systems were built which receive reflections or scattering from targets. If this reflected data is recorded, time reversed and launched into the medium again, it will focus back on the targets. This is the basis for experimental time reversal. Time reverse imaging is somewhat different in the sense that scattering from targets are recorded on the sensors, but then back propagated numerically. Narrow-band or single frequency MUSIC based time-reverse imaging algorithms have been proposed in literature for point-like targets. When this algorithm is applied to scattering from an extended target, such as a landmine, the image has good cross-range resolution, but rather poor range resolution. We propose the use of 2-D MUSIC-based algorithm to improve the near-field range resolution, which can then be used in conjunction with single frequency MUSIC to produce a final high-resolution image. A FDTD elastic-wave simulation is used to verify the results using mines and mine-like targets embedded in a heterogenous soil.

The overall objective of this work is to investigate the possibilities of using airborne IR sensors for the purpose of detecting minefield features, such as land mines. A method is proposed for temporal analysis by extracting relevant information from diurnal IR images utilizing a combination of thermodynamic modelling, signal and image processing. This paper presents results from a field test of level 2 survey in May 2003 of suspected mine-polluted areas in Croatia. Airborne data was acquired using an IR sensor mounted on a rotary wing UAV. A weather station was used to collect weather data, and pt-100 temperature sensors recorded the temperature gradient in the soil and in reference markers that were used for calibrating the IR camera. The proposed method compares simulated temporal temperature with image data collected at several times during a diurnal cycle from the same area, pixel by pixel. The images are co-registered and calibrated with respect to reference values. The numerical model is based on physical laws and is set with relevant properties, geometries, materials, surface coefficients and the influence of the actual weather sets the boundary conditions. This paper shows some results from using temporal features for detection of different relevant objects in a real minefield.

Infrared polarization has been used to investigate how well a partly covered object can be detected. It has earlier been shown that a covered object could be detected even though it was not at all possible to see the object measured in a conventional way without a polarizer. Also a measure of how much better an object is seen in polarization than without using polarization has been defined. This has earlier been applied to trip wires and partly covered surface laid personnel mines. In this work polarization measurement has been performed on realistic minefields, that has been setup by SWEDEC in Sweden. The goal has been to investigate how many mines and trip wires can be detected with polarization measurements in the LW region. The method is working well, but the present equipment are not very effective in finding mines and trip wires.

During May and June of 2003, the US Army Night Vision and Electronic Sensors Directorate (NVESD) and the Ohio State University (OSU) measured the thermal behavior of mines in an arid site. Thermistors were placed in contact with both surface-laid mines and native stones and monitored from before sunset until well after sundown. Measurements of local vegetation and measurements of the surrounding soil at 2.5 and 5 cm depths were also performed. A tripod-mounted MWIR sensor was used concurrently to collect high-resolution images to identify and understand the underlying phenomena. Data were collected during both clear, sunlit conditions and during an overcast day, but because of space limitations only data acquired under the (more typical) clear conditions are described here. The results contain a number of findings. First, local soil properties appear to have important implications for the apparent mine contrast. The same type of mine at locations only a few meters apart can show significantly different contrast with the native soil. Second, natural phenomena can be a significant clutter source. The temperature of vegetation can be similar to that of mines, and a small plant will occasionally produce a signature with a shape similar to that of a surface mine. Native stones are also a source of false alarms, but they tend to show somewhat less contrast. Third, at certain times, mines are best viewed with a low-elevation angle sensor. The construction of some mines causes the temperature of the side walls to be significantly different from that of the top surface at those times. Finally, disturbing the surface of desert soil through excavation, vehicle traffic or even repeated pedestrian traffic is often sufficient to produce a strong thermal signature. This fact could be used to advantage to detect buried mines in desert environments.

This paper describes the development and implementation of a low cost camera system that uses polarisation features of visible light for faster area reduction. The camera system will be mounted on a mechanical minefield area reduction asset, namely an AT mine roller of The HALO Trust. The automatic detection system will give an audible alarm in order to stop the AT mine roller before the rollers detonate a mine.

Electromagnetic induction sensing (EMI), between ~ 10's of Hz and 100's of kHz, may show the strongest promise for discrimination of subsurface, shallow metallic objects such as unexploded ordnance (UXO). While EMI signals penetrate the soil readily, resolution is low and responses are sometimes ambiguous. For crucial discrimination progress, maximum data diversity is desirable in terms of look angles, frequency spectrum, and full vector scattered field data. Newly developed instrumentation now offers the possibility of full vector UWB EMI data with flexible look angle and sensor distance/sweep, defined by precise laser positioning. Particulars of the equipment and resulting data are displayed. An indication is given of potential advantages for reducing the chronic ill-conditioning of inversion calculations with EMI data, when one takes advantage of the data diversity made possible by the instrumental advances. Some EMI measurement issues cannot be solved by EMI data diversity, as when small surface clutter above a much larger UXO effectively blinds an EMI sensor. EMI surveying must be supplemented by or sometimes replaced by ground penetrating radar (GPR) approaches in such instances.

Most unexploded ordinance (UXO) are heterogeneous objects containing parts of different metals, e.g., head, body, tail and fins, copper banding, etc. Recently, low frequency electromagnetic induction (EMI) sensing, based on the EM diffusion phenomena, has shown considerable progress for the detection and discrimination of UXO. EMI responses are sensitive to the type of metal (conductivity and permeability), to the distance between the sensor and scatterer, and to the coupling effects between different parts of the object. Until now, the simple dipole models used to represent EMI response have neglected the coupling and close proximity effects seen for realistic objects. These factors can interact with the particulars of excitation and observation to produce substantially varied signature patterns for a given object. This means that a key requirement in discrimination/inversion processing is to calculate very fast but very realistic EMI responses for actual target types. This work presents a new discrimination technique based on the standardized excitation approximation (SEA). The SEA seeks to identify objects in terms of their characteristic responses to sets of well defined excitations that can be used to describe any primary (excitation) field. In the new SEA system presented here, the standardized excitations are those produced by a standardized source set (SSS), in particular, fictitious magnetic sources distributed mathematically over a surface surrounding a scatterer. Several numerical results are given to illustrate the efficiency and accuracy of the proposed new technique. Finally, the spatial distribution and frequency dependence of responding equivalent sources are analyzed to demonstrate the usefulness of SSS for target discrimination.

The performance of mine detecting instruments is embedded in the behavior of a complex system. The total reliability is always composed of the intrinsic physical detection capability of the sensor, application/environmental influences and human factors. The intrinsic capability and some application factors can be investigated in laboratory measurements. Human factors, other application factors and the overall reliability, can only be evaluated in blind field trials in which the probability of detection (PoD) and false alarm rate (FAR) are measured statistically. Both of these approaches are included in CEN Workshop Agreement CWA 14747:2003, which standardizes detector testing in Humanitarian Demining. We report here the results of a study to investigate how to optimize such testing. For efficient and statistically valid field trials, the number, types and burial depths of targets, and the number of test lanes, soil types, repetitions and operators need to be carefully chosen. Laboratory results should be used to help construct field trial protocols and also to help distinguish the different contributions to the PoD and FAR, to determine where to improve insufficient performance. In this study, four models of metal detector were tested in three field trials and in the laboratory. The repeatability of the field trials is assessed, taking into account operator training and experience. Results of the laboratory tests are compared with results of the field trials and used to construct a "modular model" of the system, as used in nondestructive testing. The conclusions are, in principle, applicable to trials of other types of sensor.

Extensive studies of in-air testing of various metal detectors have been previously performed for a wide variety of targets and operating conditions. Using similar targets, we conducted a preliminary evaluation of a laboratory prototype wideband metal detector operating in the frequency domain (FD) under development at Quantum Magnetics. The wideband metal detector uses a small magnetoresistive (MR) sensor instead of an induction coil in the receive chain and can collect frequency response signatures of targets in the frequency band 100 Hz-150 kHz, thereby providing a more complete picture of a low metal mine response. These results suggest that wideband metal detection can play an important role in improving the false alarm rate (FAR) in a common detector platform by improving the amount of information provided to the fused algorithm process.

Trace detection of energetic materials has turned into a continuously growing field of interest for environmental and security reasons. The spectroscopic and imaging characterization of these materials at trace level play a major role in the development of sensing devices that enable their detection. In this work, the synthesis and imaging characterization of TNT particles over glass and gold substrates is performed. TNT solution in HPLC grade solvent was utilized as an aerosol jet coupled to a crossed beam of nitrogen gas to dry the aerosol during the path to the glass surface. Two different temperatures were set for the drier gas, 293 and 328 K. Atomic force microscopy (AFM), near field optical microscopy (NSOM), and scanning electron microscopy (SEM) reveal that a hot aerosol jet of TNT produced the smaller particles of these materials. TNT deposits resemble liquid like droplets with an ellipsoidal shape. TNT droplets prepared at room temperature and 320 K were (1736 ± 600) and (1379 ± 503)nm in diameter, respectively. The spectroscopic measurements revealed that nanosized formations correspond with TNT signature.

TNT and DNT are important explosives used as base charges of landmines and other explosive devices. They are often combined with RDX in specific explosive formulations. Their detection in vapor phase as well as in soil in contact with the explosives is important in landmine detection technology. The spectroscopic signatures of nitroaromatic compounds in neat forms: crystals, droplets, and recrystallized samples were determined by Raman Microspectroscopy (RS), Fourier Transform Infrared Microscopy (FTIR) and Fiber Optics Coupled - Fourier Transform Infrared Spectroscopy (FOC-FTIR) using a grazing angle (GA) probe. TNT exhibits a series of characteristic bands: vibrational signatures, which allow its detection in soil. The spectroscopic signature of neat TNT is dominated by strong bands about 1380 and 2970 cm^-1. The intensity and position of these bands were found remarkably different in soil samples spiked with TNT. The 1380 cm^-1 band is split into a number of bands in that region. The 2970 cm^-1 band is reduced in intensity and new bands are observed about 2880 cm^-1. The results are consistent with a different chemical environment of TNT in soil as compared to neat TNT. Interactions were found to be dependent on the physical source of the explosive. In the case of DNT-sand interactions, shifts in vibrational frequencies of the explosives as well as the substrates were found.

The photochemistry of TNT in toluene, water and methanol has been studied with femtosecond laser spectroscopy, surface reflection Fourier transform infrared absorption spectroscopy and ultraviolet-visible absorption spectroscopy measurements. Aqueous and alcoholic TNT solutions change from colorless to yellow or red upon irradiation with ultrafast 266.7 nm laser pulses. Irradiated aqueous or alcoholic TNT solutions exhibit increased absorption of light above 300 nm. Surface reflection FTIR measurements of dry deposits of irradiated samples are consistent with the formation of amines or alcohols in the photochemistry of TNT in aqueous or alcoholic solutions. In contrast, no evidence is observed in post irradiation UV-visible absorption or surface reflection FTIR measurements of TNT in toluene samples exposed to 266.7 nm femtosecond laser pulses. The results suggest that the hydroxyl group is involved in the formation of photoproducts of TNT photolysis. In addition, the results suggest that femtosecond laser photolysis is suitable for TNT detection in wet media.

Deminers and dismounted countermine engineers commonly use metal detectors, ground penetrating radar and probes to locate mines. Many modern landmines have a very low metal content, which severely limits the effectiveness of metal detectors. Canines have also been used for landmine detection for decades. Experiments have shown that canines smell the explosives which are known to leak from most types of landmines. The fact that dogs can detect landmines indicates that vapor sensing is a viable approach to landmine detection. Several groups are currently developing systems to detect landmines by “sniffing” for the ultra-trace explosive vapors above the soil. The amount of material that is available to passive vapor sensing systems is limited to no more than the vapor in equilibrium with the explosive related chemicals (ERCs) distributed in the surface soils over and near the landmine. The low equilibrium vapor pressure of TNT in the soil/atmosphere boundary layer and the limited volume of the boundary layer air imply that passive chemical vapor sensing systems require sensitivities in the picogram range, or lower. ADA is working to overcome many of the limitations of passive sampling methods, by the use of an active sampling method that employs a high-powered (1,200+ joules) strobe lamp to create a highly amplified plume of vapor and/or ERC-bearing fine particulates. Initial investigations have demonstrated that this approach can amplify the detectability of TNT by two or three orders of magnitude. This new active sampling technique could be used with any suitable explosive sensor.

Irradiating substances with pulsed neutrons results in several types of interactions which cause the emission of gamma rays. The energy of these gamma rays is characteristic of the nuclei with which the reaction occurred, and can therefore be used as an indicator of the presence of an atomic species. The PELAN system uses a pulsed neutron generator, which makes it possible to separate the gamma spectra into inelastic and capture components that are easier to interpret. Historically, the analysis of PELAN data has been based on a least squares method to extract the contribution of different elemental species present in the sample. The approach uses measured response functions for each element of interest, followed by decision rules for the identification of the materials. We have investigated an alternative approach that does not require a model and response functions. Instead, the approach determines features directly from a number of spectra of substances of interest, e.g. explosives and hazardous chemicals. The PCA method has been used to obtain indicators from the spectra. These indicators are then used for detection and identification of substances using the GLRT algorithm. The performance of the data analysis is assessed through ROC curves. A comparison of the two approaches indicates that PCA followed by GLRT technique has better performance and is more robust than the previous approach.

A comparison of the NQR parameters of the monoclinic and orthorhombic phases of TNT and their relation to the twist or dihedral angle between the plane of the NO2 substituents and that of the benzene ring as determined in the X-Ray crystal structure analysis enables an assignment of different frequencies to specific sites in the two independent molecules in the unit cell of both forms to be made. The slow transformation of the metastable orthorhombic phase to monoclinic can then be followed by monitoring the NQR spectrum in which specific lines can be assigned to molecular sites in the two phases. NQR spectra of TNT referred to in the literature often differ; this could be due partly to the TNT often being a mixture of monoclinic and orthorhombic phases and partly to changes in the spectral line width, factors which must be taken into account when NQR is used to detect landmines.

We are developing a portable, artificial olfactory system based on multiple attributes of the sense of smell to identify air-borne odors, including those associated with buried landmines. Brief (1-2 sec) air samples are drawn over an array of optically-interrogated, cross-reactive chemical sensors. These consist of polymers with high sensitivity and relatively narrow specificity for nitroaromatics (Timothy Swager, MIT), as well as those with broader responses, thus permitting discrimination among substances that may be confused for nitroaromatics. Biologically-based pattern matching algorithms automatically identify odors as one of several to which the device has been trained. In discrimination tests, after training to one concentration of 6 odors, the device gave 95% correct identification when tested at the original plus three different concentrations. Thus, as required in real world applications, the device can identify odors at multiple concentrations without explicitly training on each. In sensitivity tests, the device showed 100% detection and no false alarms for the landmine-related compound DNT at concentrations as low as 500 pp-trillion (quantified by GC/MS) - 10 times lower than average canine behavioral thresholds. To investigate landmine detection capabilities, field studies were conducted at Ft. Leonard Wood, MO. In calibration tests, signals from buried PMA1A anti-personnel landmines were clearly discriminated from background. In a limited 9 site "blind" test, PMA1A detection was 100% with false alarms of 40%. Although requiring further development, these data indicate that a device with appropriate sensors and exploiting olfactory principles can detect and discriminate low concentration vapor signatures, including those of buried landmines.

The transport of the chemical signature compounds from buried landmines in a three-dimensional minefield array has been numerically modeled using the finite-volume technique. Compounds such as trinitrotoluene and dinitrotoluene are semi-volatile and somewhat soluble in water; furthermore, they can strongly adsorb to the soil and undergo chemical and biological degradation. Consequently, the spatial and temporal distributions of such chemicals depend on the mobility of the water and gaseous phases, their molecular and mechanical diffusion, adsorption characteristics, soil water content and compaction, and environmental factors. Surface concentrations decrease, when precipitation occurs due to advective flux around the object. Deformation in the concentrations contours after rainfall is observed in the inclined surface case and it is attributed to both: the advective flux, and to the water flux at the surface caused by the inclination. The LaGrit code developed at Los Alamos National Laboratory (LANL) was used to generate the 3D grid array and to place several landmines at different underground positions. The simulations were performed by using the Finite-Element Heat and Mass-transfer code also developed originally at LANL.

A new technique for constructing score-based feature vectors for Algorithm fusion is presented. Algorithm fusion is the combining (fusing) of the outputs of multiple detection and classification (D/C) algorithms to generate a final target/nontarget decision. A 2-class D/C problem is considered: Target and Nontarget. It is assumed that multiple D/C algorithms are used to process that same raw sensor data. It is further assumed that each algorithm produces a nonnegative score for each detected object that is a measure of the degree of "target-likeness." (A score of zero is assigned to any D/C algorithm that did not detect a particular object that had been detected by another algorithm.) Several new score-based feature vectors are constructed using only the scores of the individual D/C algorithms. The feature vectors can be used as input to any feature-based classifier; for this paper, the 2-class linear classifier based on maximizing the Fisher ratio criterion has proven very effective. The different score-based feature vectors have different dimensionality. In light of Bellman's curse of dimensionality, this permits one to select the feature vector whose size is most compatible with the size of the training data set. Consequently, robust performance can be achieved.

An improved sea mine computer-aided-detection/computer-aided-classification (CAD/CAC) processing string has been developed. This robust automated processing string involves the fusion of the outputs of unique mine classification algorithms. The overall CAD/CAC processing string consists of pre-processing, adaptive clutter filtering (ACF), normalization, detection, feature extraction, optimal subset feature selection, feature orthogonalization, classification and fusion processing blocks. The range-dimension ACF is matched both to average highlight and shadow information, while also adaptively suppressing background clutter. For each detected object, features are extracted and processed through an orthogonalization transformation, enabling an efficient application of the optimal log-likelihood-ratio-test (LLRT) classification rule, in the orthogonal feature space domain. The classified objects of 4 distinct processing strings are fused using the classification confidence values as features and “M-out-of-N”, or LLRT-based fusion rules. The utility of the overall processing strings and their fusion was demonstrated with new shallow water high-resolution sonar imagery data. The processing string detection and classification parameters were tuned and the string classification performance was optimized, by appropriately selecting a subset of the original feature set. Two significant improvements were made to the CAD/CAC processing string by employing sub-image adaptive clutter filtering (SACF) and utilizing a repeated application of the subset feature selection/feature orthogonalization/LLRT classification blocks. It was shown that LLRT-based fusion of the CAD/CAC processing strings outperforms the “M-out-of-N” algorithms and results in up to a seven-fold false alarm rate reduction, compared to the best single CAD/CAC processing string results, while maintaining a high correct mine classification probability. Alternately, the fusion of the processing strings enabled correct classification of almost all mine targets, while simultaneously maintaining a very low false alarm rate. A novel investigation was also presented that illustrates and provides insights on the increased performance gains provided by utilizing LLRT-based fusion of all the different combinations of 2, 3 or 4 distinct processing strings.

In 1999 Raytheon adapted its shallow-water Side-Looking Sonar (SLS) Computer Aided Detection/Computer Aided Classification (CAD/CAC) algorithm to process side-scan sonar data obtained with the Woods Hole Oceanographic Institute's Remote Environmental Monitoring Units (REMUS) autonomous underwater vehicle (AUV). To date, Raytheon has demonstrated the ability to effectively execute mine-hunting missions with the REMUS vehicle through the fusion of its CAD/CAC algorithm with several other CAD/CAC algorithms to achieve low false alarm rates while maintaining a high probability of correct detection/classification. Mine-hunting in the very shallow water (VSW) environment poses a host of difficulties including such issues as: a higher incidence of man made clutter, significant interference due to biological sources (such as kelp or silt), the scouring of mines into the bottom, interference from surface/bottom bounce, and image distortion due to vehicle motion during image generation. These issues coupled with highly variable bottom conditions and small bottom targets make reliable hunting in the VSW environment very difficult. In order to be operationally viable, the individual CAD/CAC algorithms must demonstrate robustness over these very different mine-hunting environments. A higher normalized false alarm rate per algorithm is considered acceptable based on the false alarm reduction achieved through multi-algorithm fusion. Raytheon's recent CAD/CAC algorithm enhancements demonstrate a significant improvement in overall CAD/CAC performance across a diverse set of environments, from the relatively benign Gulf of Mexico environment to the more challenging areas off the coast of southern California containing significant biological and bottom clutter. The improvements are attributed to incorporating an image normalizer into the algorithm's pre-processing stage in conjunction with several other modifications. The algorithm enhancements resulted in an 11% increase in overall correct classification probability with an accompanying 17% reduction in false alarm rate, when averaged over the multiple environments. The paper discusses the algorithm enhancements and presents the detailed performance results.

A feature extraction method for underwater target classification is developed that exploits the linear dependence (coherence) between two sonar returns. A canonical coordinate decomposition is applied to resolve two consecutive acoustic backscattered signals into their dominant canonical coordinates. The corresponding canonical correlations are selected as features for classifying mine-like from non-mine-like objects. Test results are based on a subset of a wideband data set that has been collected at the Applied Research Lab (ARL), University of Texas (UT)-Austin. This subset includes returns from different mine-like and non-mine-like objects at several aspect angles in a smooth bottom condition. The test results demonstrate the potential of the canonical correlation-based feature extraction for underwater target classification and indicate that canonical correlation features are indeed robust to variations in aspect angle.

The work reported in this paper is focused on the particular problem of detecting mines buried in the sediment layer of shallow water environments. This is relevant to the detection of mines buried in beaches, surf zone or wet agricultural areas such as rice paddy fields. A reduced scale Electrical Impedance Tomography (EIT) detector is used in a laboratory setting to evaluate the capability of the EIT technology to detect mine-like objects buried in a layer of sand, underwater. Detection results are presented for two correlation methods developed using signatures measured for known mine-like objects located at several relative positions with respect to an electrode array. Discrimination results are also presented for two mine-like objects of similar shape and different size. Finally, recommendations are made for future research and implementation.

Registering video imagery in real-time is a demanding process. Multispectral imagery adds complexity due to the variances between different bands. This paper demonstrates a process that registers airborne multispectral imagery at a rate of thirty frames per second. It can create both mosaics and multispectral sets from a camera that captures a cycle of spectral bands with each spectral band in a separate video frame. A series of phase correlation measurements provides subpixel accuracy. Roll, pitch and yaw variances are corrected with complex polynomial interpolation.

CTI is developing a compact and efficient multispectral laser transmitter for surf zone and on-shore active imaging, under a Phase II SBIR contract with MARCORSYSCOM. The transmitter will be tested with government-furnished sensors, and is designed to search for mines and surface targets in shallow water, beach, and low vegetation areas. The system will produce range-resolved, multispectral images suitable for automatic-target-recognition (ATR) algorithm processing. CTI has performed LADAR modeling of the relevant scenario, using camera calibration data provided by NSWC. Low-risk laser components are utilized in a novel device geometry that permits the development of very high wall plug efficiency with minimal cooling requirements. Preliminary laboratory results with the transmitter subsystems are given.

Data collections were conducted using the AN/PSS-14 mine detector on three beach areas in Florida. A few samples of inert anti-tank (AT) and anti-personnel (AP) mines were buried at Jacksonville Beach, Cocoa Beach, and Clearwater Beach. The mines were buried in a variety of sand conditions varying from dry to saturated. The saturated sand conditions included the surf zone with up to two feet of water surge over the buried mine area. Test results indicate a good probability of detection (Pd) of all the buried mines by the AN/PSS-14 Ground Penetration Radar (GPR) and Metal Detector (MD), with a low false alarm rate. This paper will detail test conditions under which the mines were buried, soil dielectric and attenuation parameters measured versus water content in each condition, and interpretation of data in such highly attenuated (400-600 dB attenuation per meter) and extremely conductive soil. In addition, the theory of evanescent electromagnetic waves will be discussed in terms of the performance.

Over the past year an index has been defined which quantifies the surf zone with respect to an electro-optical (EO) system’s ability to find targets. The purpose of this index is to both normalize the EO Mine Counter Measure (MCM) systems performance expectations to the environment in which it is tested and to assess the value of its performance in an operational environment. For example, if a given system has a Probability of Detection (P_D) requirement of 90% in a clear water surf zone and is tested in murky waters the surf zone index of the murky water is used to determine what P_D is required in the murky water to yield the 90% P_D clear water requirement. The surf zone index is defined in this paper and expanded from the deterministic contrast transmittance as reported in earlier papers to a probabilistic approach. Examples of how to measure the index using readily available low cost spectral imagers such as PAR Government Systems Corporation’s Mission Adaptable Narrowband Tunable Imaging Spectrometer (MANTIS) system are given. Finally, the surf zone index usage is discussed and demonstrated.

The Airborne Littoral Reconnaissance Technologies (ALRT) project has developed and successfully demonstrated a nighttime operational minefield detection capability using commercial off-the-shelf high-power Laser Diode Arrays (LDAs). The Coastal System Station's ALRT project, under funding from the Office of Naval Research (ONR), has been designing, developing, integrating, and testing commercial arrays using a Cessna airborne platform over the last several years. This has led to the development of three test bed variants, as reported on last year: the Airborne Laser Diode Array Illuminator prototype (ALDAI-P), the original commercial array version (ALDAI-C), and the most recent wide field-of-view commercial version (ALDAI-W). Using the ALDAI-W variant because of its increased operational capabilities with higher altitudes and wider field of views, ALRT recently demonstrated nighttime operation by detecting minefields over several background variations, expanding Naval reconnaissance capabilities that had been previously limited to daytime operation. This paper describes the demonstration and shows results of the ALDAI-W test.

The objective of the Office of Naval Research (ONR) Rapid Overt Reconnaissance (ROR) program and the Airborne Littoral Reconnaissance Technologies (ALRT) project's LAMBS effort is to determine if electro-optical spectral discriminants exist that are useful for the detection of land mines in littoral regions. Statistically significant buried mine overburden and background signature data were collected over a wide spectral range (0.35 to 14 µm) to identify robust spectral features that might serve as discriminants for new airborne sensor concepts. LAMBS has expanded previously collected databases to littoral areas - primarily dry and wet sandy soils - where tidal, surf, and wind conditions can severely modify spectral signatures. At AeroSense 2003, we reported completion of three buried mine collections at an inland bay, Atlantic and Gulf of Mexico beach sites. We now report LAMBS spectral database analyses results using metrics which characterize the detection performance of general types of spectral detection algorithms. These metrics include mean contrast, spectral signal-to-clutter, covariance, information content, and spectral matched filter analyses. Detection performance of the buried land mines was analyzed with regard to burial age, background type, and environmental conditions. These analyses considered features observed due to particle size differences, surface roughness, surface moisture, and compositional differences.

We have been developing path planning techniques to look for paths that balance the utility and risk associated with different routes through a minefield. Such methods will allow a battlegroup commander to evaluate alternative route options while searching for low risk paths. Extending on previous years' efforts, we have implemented a generalized path planning framework to allow rapid evaluation and integration of new path planning algorithms. We have also implemented a version of Rapidly-Explored Random Trees (RRTs) for mine path planning which integrates path risk, path time, and dynamic and kinematic concerns. Several variants of the RRT algorithm and our existing path planning algorithms were quantitatively evaluated using the generalized path planning framework and an algorithm-dynamic evaluation graphical user interface.

We are developing a novel GPR system for stand off landmine detection. This is a bistatic GPR system, which uses a TEM horn antenna for a transmitter and a passive optical electric field sensor as a receiver. A small size passive optical sensor will be scanned on the ground surface and the acquired signal is used for synthetic aperture processing. Since the receiver is very small, it is suitable for scanning on the ground, where landmines can be buried. Fundamental laboratory test shows its capability of imaging of buried objects. Using the advantageous features of the sensor, we are testing imaging of landmines in various antenna configurations. We have adopted a newly developed wide-band system which can work up to 5GHz and discussed the experimental results.

The resistive vee dipole (RVD) loaded with the Wu-King profile has many advantages for use in ground-penetrating radar (GPR) applications. It can be designed to transmit a temporally-short pulse to a small spot on the ground. The shape of the transmitted pulse is simply related to the input signal, e.g., a derivative. The RVD also has a low radar cross section. In addition, it can be easily manufactured using a circuit board and discretely loading it with chip resistors. One drawback of the RVD is that the input impedance of the RVD increases significantly with decreasing frequency and, therefore, has a high voltage standing wave ratio (VSWR) at low frequencies, which limits the low-frequency response of the antenna. To improve the low-frequency response, a discretely-loaded resistive linear antenna (RLA) has been developed, whose basic principle of operation is the same as that of the RVD. The RLA has curved arms loaded with a modified Wu-King profile instead of straight arms loaded with the Wu-King profile. With an appropriate selection of the curve and the loading profile, the low-frequency response is significantly better for the RLA than for the RVD. The RLA has been developed using a method of moments code. The performance of the RLA is validated both numerically and experimentally.

Experiments and simulations were performed in order to assess the suitability of electrically and optically modulated scattererers (OMS) as electromagnetic field probes for measuring the pattern of ground-penetrating radar (GPR) antennas. Of special importance for the probe are its comparative performance as well as its frequency response. The former is related to the depth of modulation that the modulating device is able to reach and can be optimized with a proper choice of active element. The latter was improved by making the probe more broadband. The present work will also show the steps that have been taken to achieve better frequency response from 2 GHz to 8 GHz by means of resistively loading the probe and discuss the trade-offs involved in doing so.

The double-y balun, transitioning from a coplanar waveguide (CPW) to a coplanar strip (CPS), was originally designed for use with balanced mixers. In this paper, numerical analysis of the double-y balun is conducted using two commercial electromagnetic simulators, Momentum and HFSS. Using these numerical solvers, the effect of substrate thickness on the performance of the double-y balun is investigated. A dipole, along with the outer conductor of a coaxial feedline is modeled in NEC to illustrate the effects of an unbalanced feed on the antenna pattern of a dipole. To accurately measure the amplitude pattern of a dipole, an automated measurement system is constructed. Using this measurement system, amplitude patterns of a 3.3 GHz dipole are measured with and without the double-y balun and compared with patterns obtained via NEC.

The spatial resolution and peak signal to average noise ratio of the NIITEK ground penetrating RADAR is shown to be improved by the application of an inverse point spread function operation. Both 1-D and 2-D point spread functions are developed. The 1-D inverse PSF is developed from the step discontinuity in impedance due to the air/ground interface. The 2-D inverse PSF is developed from an image of a small sphere in air. The sequential application of the two inverse PSFs using convolution in the spatial domain compensates for two distinctly different effects. The 1-D blurring is due to AC coupling of the RADAR which produces a bipolar derivative of the narrow RADAR pulse. This is replaced by a single pulse in range which represents the location of an impedance discontinuity. The 2-D blurring is due to the wide beamwidth of the adjacent channels of the 24 channel linear array and the operation in the near field of the antenna which causes the characteristic parabolic scattering of a point signal due to adjacent channel crosstalk. Over a set of 32 landmines at a government test site an average improvement in peak signal to RMS noise ratio of 7.19 dB for 1-D only inverse filtering is realized. When 1-D processing is combined with 2-D inverse filtering, an average SNR improvement of 7.68 is realized with significant spatial resolution improvement. Subsequent processing with a moving average filter of size 2 and 4 in depth yields 8.14 and 9.88 dB net SNR improvement respectively.

Ground penetrating radar (GPR) has been shown to be useful in the detection of landmines. It is of great interest to extend this capability to discrimination between landmines and other objects cluttering the battlefield environment. Wavenumber migration processing (SAR imaging) is used here to show the ability of a GPR to determine both burial depth and size of landmines. Wavenumber migration imaging is summarized and an automated algorithm for extracting size and depth is introduced. A repeatability study is presented for ten signatures from the same metallic landmine. An example of 2D wavenumber migration imaging is presented, as well as, a summary of landmine size and depth estimates from the ten signatures.

In order to detect anti-tank mines at standoff distances, we have developed a forward-looking synthetic aperture ground penetrating radar (FLGPR). The system operates over the frequency band 766~MHz to 3.8~GHz. Our FLGPR system uses a Mill's cross transmit-receive array configuration. The receive array contains 46 Archimedean spiral antennas spaced across a 3.43 meter horizontakl aperture. The transmit aperture can be configured to contain up to 15 transmitters in one of two vertical configuations. Data is acquired as the system continuously moves forward at a speed of 2 to 8 kph. Synthetic aperture nearfield beamforming, a form of multi-look processing, is used to reduce clutter and produce significantly improved images of buried targets. Testa against actual buried mines on U.S. Army mine lanes indicate that the system can detect buried metallic and plastic anti tank mines. Images and analysis of data including blind test results are presented.

Ground Penetrating Radar (GPR) has been applied for several years to the problem of detecting both anti-personnel and anti-tank landmines. Most of the evaluation effort has focused on obtaining the end-to-end performance metrics (e.g. Pd and pfa ) of complete detection systems. This is the fourth in a series of papers in which we focus on the specific performance of one critical component of GPR systems: the antenna subsystem. In this paper, we examine several free-space characteristics of 3 prototype wideband antennas, here denoted by the terms: Resistive Vee , Antipodal Vivaldi, and Planning Systems Inc's (PSI) Archimedean Spiral antennas. Specifically, we (1) determine gain and phase properties of these antennas, (2) measure the internal reflections, (3) determine the direct coupling between antennas used in bistatic pairs, (4) measure antenna reflectivity, and (5) measure the spatial response footprints.

In this work the detection process of buried objects is presented utilizing Ground Penetrating Radar (GPR). Background removal algorithm is used to obtain the target signature and correlation process is performed to reveal the reflected target energy Then, a detection warning signal is created depending on a special process. In this work, pulsed GPR system with 1 GHz bandwith is used. Scanning speed is 0.33cm/sec in the sweeping direction and this process is repeated in the walking direction with 4 cm spatial resolution.

This paper presents the development results for three sensor technologies: metal detector (MD) array, ultrawideband (UWB) ground penetrating radar (GPR) array and biosensor sample collection and analysis system. It provides results on explosives findings for demining and demonstrates how the false alarm rate (FAR) of the MD may be reduced while maintaining high probability of detection (PD) through a data fusion (DF) system. The relevance of the results to demining and homeland security is also provided.

Hyperspectral signatures for surface mines in airborne images can have substantial variability due to the environmental conditions and subpixel mixing. Signatures are also affected by the condition of the mine. We show that a subspace representation for mine spectral properties can be used as the basis for an algorithm for subpixel mine detection that is invariant to the illumination and atmospheric conditions. A background model is estimated from the image data to support subpixel detection. The intrinsic spectral reflectance of the mine is the only input required by the algorithm. We demonstrate the performance of the algorithm for several mine types over a range of conditions and altitudes in visible through near-infrared hyperspectral images. Several of the mine types appear at a scale that is significantly smaller than a pixel.

Georgia Tech is in the second year of a Multi-University Research Initiative designed to study the impact of environmental processes on optical signatures. In particular, this program is conducting phenomenological studies on hyperspectral and polarimetric signatures of various target classes in the visible and infrared wavebands. Initial research studies have focused on landmines and the impact of various environmental factors and processes (e.g., subsurface processes) on the resultant spectral infrared signatures. A variety of approaches have been employed in this research to gain a better understanding of the impact of the environment on the spectral and polarimetric characteristics of soil and landmine signatures. These approaches include theoretical analyses, physics-based signature modeling, field measurements, and laboratory studies. Results from these studies will be presented that underscore the importance of incorporating the subsurface processes into the signature analyses. The results of these analyses have been propagated to algorithm developers to permit the creation of more robust processing techniques based on these physical analyses and models. This paper will present an overview of the program, a review of the research investigations initiated over the past year, and a summary of the results from these initial investigations.

There are currently no fielded technologies for noncontact detection of tripwires. Itres Research Ltd. and DRDC Suffield have been conducting research on optical detection of tripwires since 1996, both for hand-held and vehicle-mounted roles. A proof-of-concept brassboard imager, initially for a vehicle-mounted role, has been constructed. The imager uses a high spatial resolution, panchromatic focal plane array whose high degree of integration includes on-board digitization and flexible addressing capabilities for windowing and subimaging. Command, control and signal processing are accomplished by a computer, based on dual 1GHz Pentium III processors. Using a high level, rapid prototyping language, 1 image frame can be processed in 3 seconds. Straightforward improvements should allow true real-time operation to be achieved. Preliminary testing of the imager was conducted in the outdoor DRDC Suffield Mine Pen in January 2003. Taut, sagging and undulating tripwires of various materials were partially hidden, often nearly invisible to the naked eye, in a number of types of local vegetation. Preliminary, quasi-real-time results showed that many of the wires were detected, although a significant number of false alarms occurred. As expected with the present algorithm, sagging, undulating and highly obscured wires were often difficult to detect. The instrument, results of the trial, planned improvements and future research will be discussed.

This paper describes a polarised Short Wavelength Infra-Red (SWIR) system using Commercial Off-The-Shelf (COTS) technology, which was assessed against a variety of tripwires and backgrounds during night and day trials as part of a DSTL (UK) programme. The system comprises a polarised SWIR illuminator and cameras fitted with SWIR filters and polarisers. Various image-processing techniques were developed and evaluated including Stoke's S1 parameter, the Radon Transform and a novel and robust feature detector. Within the limits of the optical system, a tripwire recognition capability in vegetation was achieved that approached that of a human.

The UK Defence Science and Technology Laboratory (Dstl) has proven the utility of its Infrared Polarisation Sensor for forward-looking detection of flush and surface laid landmines. The system utilises a spinning polariser to analyse the polarisation content of a scene and detection is based on this analysis. This paper is based on work carried out by Dstl under the UK Applied Research Programme and focuses on an investigation into the effectiveness of applying the IPS to the detection of tripwires. The investigation and its initial results are detailed and image-processing techniques are discussed.

Detection of tripwires is an active area of investigation. Researchers at the University of Missouri and the University of Florida are jointly pursuing numerous approaches to detect both the trip wires and the mines to which they are connected. In this paper, we discuss issues related to the detection of tripwires both on a frame-by frame basis and within image sequences. A large data collection was performed under various environmental conditions. Using algorithms that operate on visible and near infrared imagery, several metrics are explored to categorize tripwire detection performance. Currently, the detection algorithm utilizes the Hough transform to detect line-like structures and then scores these candidates to distinguish between wires and linear background objects. Adaptations to the Hough transform are discussed to add robustness and to decrease the computational load. Within the sequence analysis, emphasis is placed on the use of fuzzy logic rules to integrate over time. Results of several experiments in the outdoor settings are described and analyzed.

In this paper we describe the use of a horizontally polarised Ultra-Wideband imaging radar to detect metal tripwires. A limited set of experimental measurements was processed using a wire detection algorithm designed to discriminate wires from clutter. A common threshold applied to the algorithm output successfully discriminated wires deployed at four orientations. Three independently measured cases rendered consistent results when processed using the same algorithm. These results suggest that further development is warranted to establish the performance on a statistical basis.

This paper introduces a system for landmine detection using sensor data generated by a Ground Penetrating Radar (GPR). The GPR produces a three-dimensional array of intensity values, representing a volume below the surface of the ground. First, a constant false alarm rate (CFAR) detector is used to focus attention and identify candidates that resemble mines. Next, translation invariant features are extracted by projecting the magnitude of the Fourier transformation onto the dominant eigenvectors in the training data. The training signatures are then clustered to identify prototypes. Crisp and fuzzy k-nearest neighbor rules are used to distinguish true detections from false alarms.

A characteristic of vehicle-based ground-penetrating radar is the hyperbolic signature generated by targets such as landmines. The hyperbola provides a significantly different shape from most false alarms. Here an approach is introduced that seeks to utilize all of the energy contained in this characteristic hyperbolic signature. We propose a Hyperbola Flattening Transform (HFT) that transforms hyperbolic signatures of interest into straight lines, which are in turn detected using the Radon transform. The algorithm is applied to both simulated and real data. Encouraging results are presented when applying the HFT to the problem of detecting low signal-to-noise ratio plastic mines.

The recent development of high quality sensors paired with development of advanced statistical signal processing algorithms has shown that there are sensors that can not only discriminate targets from clutter, but can also identify subsurface or obscured targets. In a previous theoretical and simulation study, we utilized this identification capability in addition to contextual information in a multi-modal adaptive algorithm where the identification capabilities of multiple sensors are utilized to modify the prior probability density functions associated with statistical models being utilized by other sensors. We assumed that the statistics describing the features associated with each sensor modality follow a Gaussian mixture density, where in many cases the individual Gaussian distributions that make up the mixture result from different target types or target classes. We utilized identification information from one sensor to modify the weights associated with the probability density functions being utilized by algorithms associated with other sensor modalities. In our simulations, this approach is shown to be improve sensor performance by reducing the overall false alarm rate. In this talk, we transition the approach from a simulation study to consider real field data collected by both handheld and vehicular based systems. We show that by appropriate modification of our statistical models to accurately match field data, improved performance can be obtained over traditional sensor fusion algorithms.

Images recorded in ground areas potentially containing surface laid land mines are considered. The first hypothesis is that the image is of clutter (grass) only, while the alternative is that the image contains a partially occluded (covered) land mine in addition to the clutter. In such a scenario, the occlusion pattern is unknown and has to be treated as a nuisance parameter. In a previous paper it was shown that deterministic treatment of the unknown occlusion pattern, in companion with the applied model, renders a substantial increase in detector performance as compared to employment of the traditional additive model. However, a deterministic assumption ignores possible correlation and additional gains could be possible by taking the spatial properties into account. In order to incorporate knowledge regarding the occlusion, the spatial distribution is characterized in terms of an underlying Markov Random Field (MRF) model. A major concern with MRF models is their complexity. Therefore, in addition to this, a less computationally demanding technique to accommodate the occlusion behavior is also proposed. The main purpose of this paper is to investigate if significant gains are possible by acknowledging the spatial dependence. Evaluation on data using real occluded targets however indicates that the gain seem to be marginal.

As part of the Commanding General of Army Material Command's Research, Development & Engineering Command (RDECOM), the U.S. Army Research Development and Engineering Center (ARDEC), Picatinny funded a joint development effort with McQ Associates, Inc. to develop an Advanced Minefield Sensor (AMS) as a technology evaluation prototype for the Anti-Personnel Landmine Alternatives (APLA) Track III program. This effort laid the fundamental groundwork of smart sensors for detection and classification of targets, identification of combatant or noncombatant, target location and tracking at and between sensors, fusion of information across targets and sensors, and automatic situation awareness to the 1st responder. The efforts have culminated in developing a performance oriented architecture meeting the requirements of size, weight, and power (SWAP). The integrated digital signal processor (DSP) paradigm is capable of computing signals from sensor modalities to extract needed information within either a 360° or fixed field of view with acceptable false alarm rate. This paper discusses the challenges in the developments of such a sensor, focusing on achieving reasonable operating ranges, achieving low power, small size and low cost, and applications for extensions of this technology.

The global landmine problem came to the attention of researchers in the mid 1990's and by 1997 several advanced and expensive sensor research programs had started. Yet, by the end of 2003, there is little sign of a major advance in the technology available to humanitarian demining programs. Given the motivation and dedication of researchers, public goodwill to support such programs, and substantial research resources devoted to the problem, it is worth asking why these programs do not seem to have had an impact on demining costs or casualty rates. Perhaps there are factors that have been overlooked. This paper reviews several research programs to gain a deeper understanding of the problem. A possible explanation is that researchers have accepted mistaken ideas on the nature of the landmine problems that need to be solved. The paper provides several examples where the realities of minefield conditions are quite different to what researchers have been led to believe. Another explanation may lie in the political and economic realities that drive the worldwide effort to eliminate landmines. Most of the resources devoted to landmine clearance programs come from humanitarian aid budgets: landmine affected countries often contribute only a small proportion because they have different priorities based on realistic risk-based assessment of needs and political views of local people. Some aid projects have been driven by the need to find a market for demining technologies rather than by user needs. Finally, there is a common misperception that costs in less developed countries are intrinsically low, reflecting low rates paid for almost all classes of skilled labour. When actual productivity is taken into account, real costs can be higher than industrialized countries. The costs of implementing technological solutions (even using simple technologies) are often significantly under-estimated. Some political decisions may have discouraged thorough investigation of cost-effective alternatives to landmine clearance.

Quadrupole resonance (QR) technology for explosives detection is of crucial importance in an increasing number of applications. For landmine detection, where the detection system cannot be adequately shielded, QR has proven to be highly effective if the QR sensor is not exposed to radio frequency interference (RFI). However, strong non-Gaussian RFI in the field is unavoidable, making RFI mitigation a critical part of the signal processing. In this paper, a statistical model of the non-Gaussian RFI is presented. The QR model is used within the context of an adaptive filtering methodology to mitigate RFI, and this approach is compared to other RFI mitigation techniques. Results obtained using both simulated and measured QR data are presented.

The quadrupole resonance (QR) technology can be used as a confirming sensor for buried plastic landmine detection by detecting the explosives (e.g., TNT and RDX) within the mine. We focus herein on the detection of TNT via the QR sensor. Since the frequency of the QR signal is located within the AM radio frequency band, the QR signal can be corrupted by strong radio frequency interferences (RFIs). Hence to detect the very weak QR signal, RFI mitigation is essential. Reference antennas, which receive RFIs only, can be used together with the main antenna, which receives both the QR signal and the RFIs, for RFI mitigation. By taking advantage of the spatial correlation of the RFIs received by the antenna array, the RFIs can be reduced significantly. However, the RFIs are usually colored both spatially and temporally and hence exploiting only the spatial diversity of the antenna array may not give the best performance. We exploit herein both the spatial and temporal correlation of the RFIs to improve the TNT detection performance. First, we consider exploiting the spatial correlation of the RFIs only and propose a maximum likelihood (ML) estimator for parameter estimation and a constant false alarm rate (CFAR) detector for TNT detection. Second, we adopt a multichannel autoregressive model to take into account the temporal correlation of the RFIs and devise a detector based on the model. Third, we take advantage of the temporal correlation by using a two-dimensional robust Capon beamformer (RCB) with the ML estimator for improved RFI mitigation. Finally, we combine the merits of all of the three aforementioned approaches for TNT detection. The effectiveness of the combined method is demonstrated using the experimental data collected by Quantum Magnetics, Inc.

With current signal processing techniques, successful discrimination between UXO (Unexploded Ordnance) and clutter depends on characteristics that are consistent across all examples of an ordnance type. Real UXO, however, exhibit many differences from instance to instance, such as varying degrees of damage sustained, degradation over time, orientation in the ground, and even differences in design. Thus, a given ordnance type, such as 60mm shells, will exhibit a wide range of signal responses, making it difficult to distinguish these items from clutter unless these fundamental differences are taken into account using appropriate mathematics. This paper will examine optimal methods of using frequency-domain analysis of wideband electromagnetic induction (EMI) for detection.

Current idealized forward models for electromagnetic induction (EMI) response can be defeated by the characteristic material and geometrical heterogeneity of realistic unexploded ordnance (UXO). A new, physically complete modeling system includes all effects of these heterogeneities and their interactions within the object, in both near and far fields. The model is fast enough for implementation in inversion processing algorithms. A method is demonstrated for deriving the model parameters from straight forward processing of training data from a defined measurement protocol. Depending on the EMI sensor used for measurements, the process of inferring model parameters is more or less ill-posed. More complete data can alleviate the problem. For a given set of training data, special numerical treatment is introduced to take the best advantage of the data and obtain reliable model parameters. This fast model is implemented in a "fingerprint" testing approach in which two different UXOs are identified from the measurement data. Preliminary results showed that this fast model is promising for UXO identification.

The handheld F3 metal detector, developed by the MineLab Corporation, measures the responses of buried objects to electromagnetic pulses. These responses can be processed to determine whether a landmine is present. The simplest processor calculates the total energy in the response, thereby reducing the entire spatial and temporal response to a single value. This value, proportional to the amount of metal in the object, can then be compared to a predetermined threshold. The drawback of this common approach is that, although the threshold may be set so that few, if any, mines are missed, doing so may result in a high false alarm rate. Previous work has demonstrated that incorporating physics-based features into a Bayesian detection framework and performing simple, one-dimensional regional processing can significantly reduce the false alarm rate while maintaining the desired level of detection. Based on these promising results, this approach has been extended to incorporate two-dimensional regional processing. At the test site, data was collected both manually and robotically using nearly identical protocols. Thus, in theory, measured responses should be similar and algorithm performance equivalent whether the detector was operated by a robot or a human. The robustness of various algorithms was evaluated by comparing performance across manual and robotic data sets. Certain physics-based feature detectors were relatively unaffected by the response variability introduced unintentionally by the human operator. However, other algorithms that incorporate more sensitive, often regional, features were able to provide greater gains for the robotic data set than for the manual data set. These results imply that there may be a tradeoff between performance and practical issues that need to be addressed when selecting an algorithm for implementation in a field setting.

The focus of this study is the evaluation of emerging soil moisture models as they apply to infrared, radar, and acoustic sensors within the scope of countermine operations. Physical, chemical, and biological processes changing the signature of the ground are considered. The available models were not run in-house, but were evaluated by the theory by which they were constructed and the supporting documentation. The study was conducted between September and October of 2003 and represents a subset of existing models. The objective was to identify those models suited for simulation, define the general constraints of the models, and summarize the emerging functionalities which would support sensor modeling for mine detection.

Previous modeling studies and experimental work have demonstrated that soil physical properties have a significant effect on most sensors for the detection of buried land mines. While a modeling approach allows for testing of the effects of a wide range of soil variables, most experimental work is limited to (field) soils with poorly known or controlled properties. With this in mind, we constructed a new outdoor test site with full control of soil water content and continuous monitoring of important soil properties and environmental conditions. In three wooden frames of 2 x 2 x 1 meter, filled with different soil types (sand, loam, and clay), we buried low-metal anti-tank and anti-personnel land mine simulants. We installed time domain reflectometry (for measurement of soil water content) and temperature probes at different depths above and below the land mines as well as in homogeneous soil away from the land mines. In this paper we document the features of this new test site and present results from the monitoring equipment.

Attenuated total reflection (ATR) spectroscopy is a well established optical technique investigating fundamental molecular vibrations in the mid-infrared (MIR) spectral regime for a wide variety of samples including liquids, thin films and powders. In the present study, first results simulating the influence of weathering processes on the spectral characteristics of soils are discussed. In particular, the effect of wetting and drying cycles on IR spectra of fine quartz (SiO₂) powders has been investigated with ATR techniques. Resulting from a wetting and drying cycle, the sample spectra of quartz powders revealed significantly increased absorption intensities throughout the spectral region of interest (1400-600 cm^-1). We hypothesize that this effect results from a higher packing density of the particles following the wetting procedure with the fines packed into interstitial spaces closer to the ATR waveguide surface. Moreover, a strong red shift of approx. 40 cm^-1 of the absorption band assigned to asymmetric SiO₄ stretching vibrations (1050 cm^-1 to 1250 cm^-1) could be observed. Both effects, increase in intensity and spectral shift, are reversed by mechanically disturbing the cemented powder after the wetting/drying cycle. Experiments with s- and p-polarized infrared radiation show similar (reversible) spectral shifts for this particular frequency range. It is expected that these findings will lead to better understanding of the spectral characteristics of soil in the mid-infrared spectral domain providing improved interpretation of data retrieved from disturbed soils e.g. potential landmine sites during hyperspectral imaging.

In this paper we present results from our research on the impact of various soil types and processes on LWIR broadbandand hyperspectral signatures of landmines and the surrounding soil. These analyses utilized our digital hyperspectral models of the combined landmine-soil system for the computation of these signatures. These physics-based models incorporate models for external environmental processes and allow soil thermal parameters to be set as a function of the subsurface conditions (i.e., porosity, moisture content). Under this research effort, signature computations were conducted for various soil types as a function of the underlying soil conditions in order to examine the relative impact of these conditions on both the broadband and spectral LWIR signatures. Of particular interest were changes in spectral features and contrast changes due to soil water content. Results from the digital signature computations will be presented along with an analysis of the signature features. A comparison of the digital signature calculations to measured data will be included in this discussion. A brief description of the signature model will also be presented.

Thermal signatures of buried land mines depend on a complex combination of environmental conditions, soil properties, and properties and burial depth of the land mine. Due to the complex nature of the problem most modeling and experimental efforts to understand thermal signatures of land mines have focused on the effects of one or a few variables. Of these variables, the effect of wind speed has received little attention in modeling and experimental studies. In this contribution we discuss the role of wind in the generation of thermal images and we present results of field experiments at the outdoor land mine detection test facility at New Mexico Tech. Here, several anti-tank and anti-personnel land mine simulants have been buried in sand, loam, and clay soils. During the measurements the environmental and soil conditions were continuously monitored using a fully equipped weather station and using probes for measurements of soil temperature and soil water content.

Newer detectors are growing in capability to discriminate those signals measured over mines from those signals that can be causally related to local variations in the soil. Monitoring and measuring the key properties governing these local variations are being looked at increasingly as a means to predict performance measures for given detectors, as well as to counter the occurrence of such signals in an effort to minimize false alarms. Currently, an ongoing government research project working to develop enhancements to the Handheld Standoff Mine Detection System (HSTAMIDS) technology resulted in a series of data collections acquired in four different types of soil environments: 1) temperate/loamy, 2) temperate/grassy/gravel, 3) arid/gravel/sand, and 4) tropical/laterite. At each of these locations, data was collected using the HSTAMIDS technology to provide a range of environmental conditions against which the performance of this handheld detector could be assessed. This project is obtaining similar electrical and magnetic measurements in these areas to use these measurements to monitor any changes in detection performance that might be introduced due to local soil variations, as well as to provide a preliminary estimate of the robustness of future HSTAMIDS detection enhancements across a variety of environments.

The presence of magnetic iron oxides in the soil can seriously hamper the performance of electromagnetic sensors for the detection of buried land mines and unexploded ordnance (UXO). Previous work has shown that spatial variability in soil water content and texture affects the performance of ground penetrating radar and thermal sensors for land mine detection. In this paper we aim to study the spatial variability of iron oxides in tropical soils and the possible effect on electromagnetic induction sensors for buried low-metal land mine and UXO detection. We selected field sites in Panama, Hawaii, and Ghana. Along several horizontal transects in Panama and Hawaii we took closely spaced magnetic susceptibility readings using Bartington MS2D and MS2F sensors. In addition to the field measurements, we took soil samples from the selected sites for laboratory measurements of dual frequency magnetic susceptibility and textural characteristics of the material. The magnetic susceptibility values show a significant spatial variation in susceptibility and are comparable to values reported to hamper the operation of metal detectors in parts of Africa and Asia. The absolute values of susceptibility do not correlate with both frequency dependence and total iron content, which is an indication of the presence of different types of iron oxides in the studied material.

Some soils can adversely affect the operation of sensitive metal detectors widely used to detect buried landmines. Although there has been some related work in geophysics, researchers in metal detection techniques, until very recently, seem to have largely ignored the issue of problem soil. As a result, rigorous scientific investigations of how soil electromagnetic properties may affect the operation of metal detectors are lacking. Thus, there is a need for theoretical and experimental studies to clarify which electromagnetic properties are important and to what extent they affect the performance of metal detectors of various designs. The paper presents a systematic analytical framework, based on existing work in geophysics and non-destructive testing, for studying the effects of soil electromagnetic properties on the functioning of metal detectors. For this initial study the burial medium is modelled as a half-space. While soil electrical conductivity has been assumed to be real and independent of frequency, soil magnetic susceptibility has been modelled as complex and frequency dependent. Simplified versions of the analysis techniques have been applied to three selected cases of practical importance, namely, non-conducting soil with constant susceptibility, non-conducting soil with frequency-dependent susceptibility and non-magnetic soil with constant conductivity. Results from a preliminary analysis of even these simple cases explain a number of previous experimental observations, for example, the greater influence of magnetic properties than of electrical conductivity on the performance of metal detectors.

Physical properties, such as soil moisture, magnetic susceptibility and electrical conductivity (EC) are sources of signal interference for many landmine detectors. Soil EC mechanisms and their relationship to moisture are being studied to increase the soil EC prediction accuracy by radar remote sensing, airborne and ground electromagnetic (EM) methods. This is required for effective detection operations in problematic regions of the world. Results indicate that responses of free water and bound water to drying rates and EC are very different, to the extent that moist clay-poor soil may have lower EC compared to dryer clay-rich soil at certain moisture contents. These suggest that soil EC prediction should start with analyses of radar remote sensing data acquired on separate days, followed by high frequency airborne EM surveys, and validation by ground EM surveys and laboratory soil sample analyses. Due to the various expertise required, a team of relevant experts (e.g., geology, geophysics, remote sensing, petrophysics, agriculture, soil physics, electrical engineering and demining) should be organized to provide information on detector viability for demining in problematic areas in the world. It is also proposed to develop wide frequency band EM systems to provide much of the required information in one measurement.

In previous work we have shown that GPR signatures are affected by soil texture and soil water content. In this contribution we will use a three dimensional electromagnetic model and a hydrological soil model to explore in more detail the relationships between GPR signatures, soil physical conditions and GPR detection performance. First, we will use the HYDRUS2D hydrological model to calculate a soil water content distribution around a land-mine. This model has been verified against measured soil water distributions in previous work. Next, we will use existing pedotransfer functions (e.g. Topp, Peplinski, Dobson, Ulaby) to convert the predicted soil water contents around the land-mines as well as known soil textures and bulk densities into soil parameters relevant to the electromagnetic behaviour of the soil medium. This will enable a mapping between the hydrological model and the electromagnetic GPR model. Using existing and new laboratory and field measurements from the land-mine test facilities at TNO-FEL we will make a first attempt to verify our modelling approach for the prediction of GPR signatures in field soils. Finally a detection algorithm is used to evaluate the GPR detection performance with respect to changing environmental soil conditions.

Under a current Army Research Office Multi-University Research Initiative, Georgia Tech has engaged in the development of detailed environmental models for studying the impact of various soil and environmental processes on landmine signatures. Under this effort, several landmine and false target physics based models have been developed; these models cover the three general cases for landmines: buried, flush-buried, and surface landmines. Each model was developed in concert with a corresponding soil model; specifically, a detailed thermal model of a landmine was incorporated within a correspondingly detailed thermal model of the soil. These models incorporate high spatial resolution (≤ 5.5 cm), three-dimensional heat transfer mechanisms, and environmental influences (i.e., meteorological information, subsurface processes) into the signature computations. This paper will present a detailed description of the models, the environmental processes currently incorporated and example signature results for both LWIR broadband and hyperspectral sensors. The representation of various soil processes in the models will also be included in the discussion.

Extremely low frequency measurements, below 30 Hz, of solid, thin-, and, thick-walled steel (permeable) cylinders with length-to-diameter ratios of approximately 4 are described and compared with the predicted response computed using a frequency domain finite element method (FDFEM). Measurements were made using a conventional EMI test setup consisting of a Hewlett Packard 89410 vector signal analyzer, rectangular transmitting and a figure-eight (bucked) receiving coil, along with appropriate transmitter and receiver coil amplifiers. All cylinders were measured with the predominant component of the excitatory magnetic field both aligned with and orthogonal to (two distinct measurements) the cylinder's axis. Measurements were made with and without a centered copper ring on the cylinders. The ring simulates the so-called rotating bands on actual UXO. Not surprisingly, we observed that the quadrature peak of the response shifts down in frequency much more when the axis of the ringed cylinder is aligned with the excitatory magnetic field than when perpendicular to it. Our measurements indicated that the real part of the response of the smallest cylinders measured asymptotically approaches its DC value around 1 Hz while the largest of the cylinders measured does not asymptote until well below 1 Hz. It appears that target information that may be crucial for discrimination purposes, especially for larger targets, exists at frequencies well below 30 Hz. Extremely low frequency measurements, especially with data averaging (stacking), can be a rather time consuming process, and therefore it is not likely that such measurements can be made from a moving platform. However, once an object of interest has been detected, the target can be reacquired and the measurement taken with the sensor stationary with respect to the target (sometimes referred to as a qued approach). As our measurements and simulations indicate, the qued method may be necessary if large solid UXO are to be distinguished from large thin-walled clutter objects.

Predicting the degree of burial of mines in soft sediments is one of the main concerns of Naval Mine CounterMeasures (MCM) operations. This is a difficult problem to solve due to uncertainties and variability of the sediment parameters (i.e., density and shear strength) and of the mine state at contact with the seafloor (i.e., vertical and horizontal velocity, angular rotation rate, and pitch angle at the mudline). A stochastic approach is proposed in this paper to better incorporate the dynamic nature of free-falling cylindrical mines in the modeling of impact burial. The orientation, trajectory and velocity of cylindrical mines, after about 4 meters free-fall in the water column, are very strongly influenced by boundary layer effects causing quite chaotic behavior. The model's convolution of the uncertainty through its nonlinearity is addressed by employing Monte Carlo simulations. Finally a risk analysis based on the probability of encountering an undetectable mine is performed.

We consider landmine detection using forward-looking ground penetrating radar (FLGPR), which is quite challenging due to the weak signal returns of landmines. The two main challenging tasks include extracting intricate structures of the target signals from the radar imagery and adapting the classifier to the surrounding environment through learning. Through the time-frequency analysis, we find that the most discriminant information is time-frequency localized. This observation motivates us to use the wavelet packet transform to sparsely represent the signals with the discriminant information encoded into several bases. Then the sequential floating forward selection method is used to extract these components and thereby a neural network classifier is designed. To further improve the classification performance, the AdaBoost algorithm is used. We modify the original AdaBoost algorithm to integrate the feature selection process into each iteration. Experimental results based on measured FLGPR data are presented, showing that with the proposed classifier, a significant improvement on both the training and the testing performances can be achieved.

A system is under development at the Georgia Institute of Technology that utilizes a seismic source to propagate Rayleigh waves through a medium such as soil. Non-surface-contacting electromagnetic sensors are used to detect the displacement of the medium created by interaction of the Rayleigh waves with a target, such as a landmine. The system has been tested in a relatively uncluttered medium and has yielded encouraging results, demonstrating that the system is effective for the detection of targets buried just below the surface. The system performs well in an uncluttered medium. However, when the medium is filled with a large number of scattering objects, the Rayleigh wave will be broken up by the scatterers in the medium to the point that the wave front no longer interacts with the target as it would in an uncluttered medium. This causes detection of a target to be uncertain or impossible. In an effort to extend the application of this system to a highly cluttered medium, the time reversal method is applied to the seismic system, and evaluated for focusing Rayleigh wave fronts at a desired location. Numerical and experimental results are presented for a propagation medium with no scatterers present, and with multiple scatterers present. Time-reverse focusing results are also compared to uniform excitation and time-delay beamforming methods.

This paper presents some advancement in the detection algorithms using EMI sensor, GPR sensor and their fusion. In the EMI algorithm, we propose the application of the weighted distributed density (WDD) functions on the wavelet domain and the time domain of the EMI data for feature based detection. A multilayer perceptron technique is then applied to discriminate between mine and clutter objects based on the wavelet domain and time domain features separately. When the results from the two domains are fused together, the probability of false alarms is reduced by a factor of two. The enhancement in the GPR algorithm includes the depth processing which selects a certain data segment below the ground surface for detection, as well as utilizing the phase variation of the signal return across a mine to achieve better detection. Finally, we present fusion results from EMI and GPR sensors to demonstrate that the two sensors provide complementary information and when they are properly fused together the probability of false alarm can be reduced significantly.

CyTerra's dual sensor HSTAMIDS system has demonstrated exceptional landmine detection capabilities in extensive government-run field tests. Further optimization of the highly successful PentAD-class algorithms for Humanitarian Demining (HD) use (to enhance detection (Pd) and to lower the false alarm rate (FAR)) may be possible. PentAD contains several input parameters, making such optimization computationally intensive. Genetic algorithm techniques, which formerly provided substantial improvement in the detection performance of the metal detector sensor algorithm alone, have been applied to optimize the numerical values of the dual-sensor algorithm parameters. Genetic algorithm techniques have also been applied to choose among several sub-models and fusion techniques to potentially train the HSTAMIDS HD system in new ways. In this presentation we discuss the performance of the resulting algorithm as applied to field data.

Outstanding landmine detection has been achieved by the Handheld Standoff Mine Detection System (HSTAMIDS system) in government-run field tests. The use of anomaly detection using principal component analysis (PCA) on the return of ground penetrating radar (GPR) coupled with metal detection is the key to the success of the HSTAMIDS-like system algorithms. Indications of nonlinearities and asymmetries in Humanitarian Demining (HD) data point to modifications to the current PCA algorithm that might prove beneficial. Asymmetries in the distribution of PCA projections of field data have been quantified in Humanitarian Demining (HD) data. An initial correction for the observed asymmetries has improved the False Alarm Rate (FAR) on this data.

The assumption is that removal of elements of clutter from the frequency stepped ground penetration radar (GPR) signal data will improve the performance of any detection algorithms. Clutter comes in the form of internal system interference, cross-coupling signals between antennas, and soil artifacts (soil layers, rocks, non-homogeneous material, grass, etc.). The assumption is that the frequency stepped radar has a number of steps that cover a fixed bandwidth, and that the radar is phase coherent from step to step and over time. Processing consists of transforming the signal data into the spatial-frequency dimension and applying a set of filters, and then transforming into the range (bandwidth compression) dimension. The developed filters remove spectral components that are associated with signal returns from clutter elements. Examples using data from the US Army AN/PSS-14 mine detection system operating over inert mines are presented.

A new preprocessing and feature extracting approach for classification of non-metallic buried objects are aimed using GPR B-scan data. A frequency-domain adaptive filter without a reference channel effectively removes the background signal resulting mostly from the discontinuity on the air-to-ground path of the electromagnetic waves. The filter only needs average of the first five A-scans as the reference signal for this elimination, and also serves for masking of the B-scan in the frequency-domain. A preprocessed GPR data with significantly suppressed clutter is then obtained by precisely positioning the Hanning window in the frequency-domain. A directional correlation function defined over a B-scan frame gives distinctive curves of buried objects. The main axis of directional correlation, on which the pivotal correlating pixels and short lines of pixels being correlated are considered, makes an angle to the scanning direction of the B-scan. This form of correlation is applied to the frame from the left-hand and the right-hand side and two over-plotted curves are obtained. Nine measures as features emphasizing directional signatures are extracted from these curves. Nine-element feature vectors are applied to the two-layer Artificial Neural Network and preliminary results over test set are promising to continue to comprehensive training and testing processes.

Two new features are presented to improve the detection of Anti-Tank (AT) landmines using Ground Penetrating Radar (GPR). A simplified three dimensial physics based model is used as the basis for the features. We combine these features with the results of an algorithm known as LMS. We present promising feature detection algorithms known as Rings N' Things (RNT) and Cross Diagonal Enhancement Processing (CDEP) and our approach to combining the new features with the LMS features using logistic regression techniques. Test results from data gathered at multiple sites covering hundreds of mines and thousands of square meters is analyzed and presented.

Designing robust landmine detection algorithms for ground penetrating radar (GPR) remains a challenging task due to variations of environmental conditions and diverse clutter objects in the soil, among others. The problem is aggravated for handheld systems by introducing operator motion and by the position uncertainty. Even though aggregating consecutive GPR samples to form multi-sample features seems to be an intuitively sensible approach to improve Pd/Pf, determining multi-sample features that are robust to the operator motion and position uncertainty is a formidable task. In this paper, we propose an ATR method to identify mines based on handheld GPR data collected for regional processing, where systematic operator motion is required and perhaps some position information is collected along with the data. The regional processing is intended to be conducted after other initial detection methods have identified an area for further interrogation. In this study, we will use GPR data that were collected by a robotic arm. In order for the developed ATR method to be applicable to data collected by human operators, which have greater position uncertainty, we focus on features that can still be used either directly or with minor modification when accurate sensor positions are not available. We tested two classes of classifiers, Support Vector Machines (SVM) and Gaussian Mixtures (GM). For both classifiers, less complex forms of the classifiers outperform those with more complicated structures. The reason is that the training set is relatively small compared to the diversity of the mines and the clutter objects in the training set.

An analysis of the utility of region-based processing of Ground Penetrating Radar (GPR) and Electromagnetic Induction (EMI) is presented. Algorithms for re-sampling GPR data acquired over non-rectangular and non-regular grids are presented. Depth-dependent whitening is used to form GPR images as functions of depth bins. Shape, size, and contrast-based features are used to distinguish mines from non-mines. The processing is compared to point-wise processing of the same data. Comparisons are made to GPR data collected by machine and by humans. Evaluations are performed on calibration data, for which the ground truth is known to the algorithm developers, and blind data, for which the ground truth is not known to the algorithm developers.

This paper addresses the problem of detecting anti-personnel landmines using features extracted via the fractional Fourier transform. A preliminary clutter reduction system is used to reduce the effects of ground clutter. The processed signatures are then examined using the fractional Fourier transform, where pertinent target features are extracted. These features are then presented to a nonparametric landmine detection system. The focus is on identifying and using fractions of the Fourier transform that will lead to better detection performance assuming that the landmine is within reasonable depth variation below the surface. The performance of the proposed detection scheme is assessed in terms of detection rate and false alarm rate using real landmine signatures. The effects of removing ground clutter are examined, and the effects of knowing the mine depth a priori are also discussed. The manner in which radar scans are processed via fractional Fourier is also discussed.

An automated methodology for combining Ground Penetrating Radar features from different depths is presented and analyzed. GPR data from the NIITEK system are processed by a depth-dependent, adaptive whitening algorithm. Shape and contrast features, including compactness, solidity, eccentricity, and relative area are computed at the different depths. These features must be combined to make a decision as to the presence of a landmine at a specific location. Since many of the depths contain no useful information and the depths of the mines are unknown, a strategy based on sorting is used. In a previous work, sorted features were combined via a rule-based system. In the current paper, an automated algorithm that builds a decision rule from sets of Ordered Weighted Average (OWA) operators is described. The OWA operator sorts the feature values, weights them, and performs a weighted sum of the sorted values, resulting in a nonlinear combination of the feature values. The weights of the OWA operators are trained off-line in combination with those of a decision-making network and held fixed during testing. The combination of OWA operators and decision-making network is called a FOWA network. The FOWA network is compared to the rule-based method on real data taken from multiple collections at two outdoor test sites.

This paper describes the further development of the MINETECT affordable humanitarian mine detector produced by ERA Technology with sponsorship from the UK Department for International Development. Using a radically different patented approach from conventional ground penetrating radar (GPR) designs in terms of the man machine interface, MINETECT offers simplicity of use and affordability, both key factors in humanitarian demining operations. Following trials in 2002 and reported at SPIE 2002, further development work including research on classifying mines, based on data from planned trials in the United Kingdom, is presented. MINETECT has the capability of detecting completely non-metallic mines and offers a considerable improvement in hand-held mine detection.

Recent success in using a laser Doppler vibrometer (LDV) based acoustic-to-seismic (A/S) landmine detection [Sabatier, J. M. and Xiang, N. IEEE Trans. Geoscience and Remote Sensing 39, 2001, pp.1146-1154.; Xiang, N. and Sabatier, J. M., J. Acoust. Soc. Am. 113, 2003, pp. 1333-1341] and a ground penetrating synthetic aperture radar (GPSAR) [Bradley et al. Proc. SPIE, 4038, pp.1001-1007, 2000] suggested a novel configuration of fused sensors comprised of a LDV-based A/S detection sensor and a GPSAR. Extensive field experiments revealed that these two technologies can be considered 'orthogonal'. When used in concert, a fused configuration may significantly improve the probability of detection and reduce the false alarm rate. They function best against different types of landmines under different burial conditions because they exploit disparate phenomena to detect mines. In order to better understand the fused detection ability, a co-located field experiment has been conducted using both a LDV-based A/S sensor and a GPSAR. This paper will discuss the comparative experimental study using the recent co-located field scanning results.

In highly contaminated unexploded ordnance (UXO) cleanup sites, multiple metallic subsurface objects may appear within the field of view of the sensor simultaneously, both for electromagnetic induction (EMI) and ground penetrating radar (GPR). Sensor measurements consist of an a priori unknown mixture of the objects' responses. The two sensing systems can provide different kinds of information, which are complementary and could together produce enhanced UXO discrimination in such cases. GPR can indicate the number of objects and their approximate locations and orientations. This data can then serve as prior information in EMI modeling based on the standardized excitation approximation (SEA). The method is capable of producing very fast, ultra-high fidelity renderings of each object’s response, including all effects of near and far field observation, non-uniform excitation, geometrical and material heterogeneity, and internal interactions. Given good position information, the SEA formulation inverts successfully for EMI parameters for each of the two objects, using EMI data in which their signals overlap. The values of the inferred parameters, in terms of their frequency and spatial patterns for an object's response to each basic excitation, are unique characteristics of the object and could thus serve as a basis for classification.

To support the development of advanced algorithms for hand-held detectors, it is desirable to collect data with a specific sweep rate, height and spacing. In addition, it is also important that the position of each data point produced by the detector is known. Since it is impossible for a human operator to precisely control these sweep parameters, we have developed a semi-autonomous robotic data collection system. It is designed as a portable robot with a 2-axis manipulator that can be used to sweep any hand-held detector at a precise sweep rate, height, and spacing. It is also equipped with an interface to the hand-held detector, so it can log the output data during the sweeping motion. It also tags the output data with the position data from the on-board positioning system. As a result, we can construct an accurate 2-D or 3-D grid of the detector's output as a function of horizontal and vertical position of the detector. The manipulator is also equipped with force sensing capability that can be used to sense terrain height or collision. To increase deployment flexibility, all functions of the robot are controlled through a wireless communication link by a hand-held computer with a maximum operating distance of at least 100m. Through the hand-held computer, the operator can move the robot, and program its behavior using a script based motion sequencer. The robot has been deployed successfully on several data acquisition activities, and successfully produced high-resolution 2-D map of the buried targets.

An overpass solution for antitank pressure-fused mines was developed and demonstrated for a teleoperated, four-wheeled experimental unmanned ground vehicle hosting a mine detection system using a downlooking ground-penetrating radar. The capability to overpass pressure-activated antitank mines is one way to protect the vehicle. The requirement was to make the vehicle overpass capable by giving it an average footprint pressure of no more than 5 psi using commercially available equipment and without requiring any vehicle modification. An overpass solution was developed and demonstrated using low-pressure, minimal casing rigidity tires that produce a uniformly low ground pressure and enable the vehicle to exert less than the minimum force required to activate the large majority of pressure-fused antitank mines. Overpass requirements are discussed in terms of antitank mine threats, pressure plate size, activation forces, and ground pressure distribution uniformity. A variable-load tire footprint pressure measurement system and laboratory were developed and laboratory evaluation of a number of tire candidates was completed. Laboratory results were demonstrated through field performance demos of the selected low-pressure tire. Results present the successful overpass of various threat representative antitank mines with the pressure plate elevations/exposures at various positions relative to local grade.

Anti-tank (AT) landmines slow down and endanger military advances and present sizeable humanitarian problems. The remediation of these mines by direct human intervention is both dangerous and costly. The Intelligent Systems & Robotics Group (ISRG) at New Mexico Tech has provided a partial solution to this problem by developing an Unmanned Ground Vehicle (UGV) to remediate these mines without endangering human lives. This paper presents an overview of the design and operation of this UGV. Current results and future work are also described herein. To initiate the remediation process the UGV is given the GPS coordinates of previously detected landmines. Once the UGV autonomously navigates to an acceptable proximity of the landmine, a remote operator acquires control over a wireless network link using a joystick on a base station. Utilizing two cameras mounted on the UGV, the operator is able to accurately position the UGV directly over the landmine. The UGV houses a self-contained drill system equipped with its own processing resources, sensors, and actuators. The drill system deploys a neutralizing device over the landmine to neutralize it. One such device, developed by Science Applications International Corporation (SAIC), employs incendiary materials to melt through the container of the landmine and slowly burn the explosive material, thereby safely and remotely disabling the landmine.

A government-funded effort was initiated to further develop algorithms based on the technology used in the U.S. Army’s latest handheld standoff mine detection system (HSTAMIDS). To this end, a complete multisensor (EM/GPR) baseline signature data set was acquired in the spring of 2003 over targets of interest for landmine detection. These were provided at a government-run test site in the eastern U.S. where hundreds of buried inert mines and discrete clutter objects are available for such signature measurements. Bringing the HSTAMIDS detector technology to this site, in conjunction with a tethered data acquisition hardware and platforms, resulted in a complete baseline multisensor signature data collection. Due to the multisensor nature of the HSTAMIDS technology, the properties of this data collection include total and real-time collocation of electromagnetic and radar sensors. Processed examples of signatures of objects of interest from this baseline signature data set are presented here, along with a summary of the use to which this data set has been put so far. The means for future requests for access to the baseline data set by individual researchers for further algorithm work are also detailed.

In the past decade, two portable systems for the discrimination of unexploded ordnance (UXO) have been developed: the PINS system and the PELAN system. While technically portable, each of these systems has ancillary equipment and wires which make them cumbersome for explosive ordnance disposal (EOD) personnel. Also, moving these systems from place to place is time-consuming. We have developed a mobile platform called the Field Utility Vehicle (FUV) to mitigate these burdens. The FUV provides power and communications for these devices, is self-propelled, and has a powered lift mechanism to adjust the height of these systems to any shell size. The FUV concept originated during a demonstration of the PELAN system at the White Oak Naval Surface Warfare Center conducted by US Navy EOD Technology Division personnel. Currently, a prototype FUV has been built and tested at the Western Kentucky University’s Applied Physics Institute. The results of these tests and cost-benefit of employing the FUV in UXO clean-up operations will be shown.

Hyperspectral imaging is an important technology for the detection of surface and buried land mines from an airborne platform. For this reason, hyperspectral was included with SAR sensors in the two deployments that were executed by the CECOM RDEC Night Vision and Electronic Systems Directorate (NVESD) in Fall 2002 and in Spring 2003. The purpose of these deployments was to bring together a wide variety of airborne sensors for the detection of mines, with well ground-truthed targets. The hyperspectral sensors included the Airborne Hyperspectral Imager (AHI), a University of Hawaii LWIR HSI sensor and the Compact Airborne Spectral Sensor (COMPASS), an NVESD VNIR/SWIR sensor. Both a high frequency SAR and a ground penetrating radar were also flown. These experiments were carried out at sites where an extensive array of buried and surface mines were deployed. At the first location, on the east coast, the mines were deployed against several different backgrounds ranging from bare dirt to long grass. At the second location in the desert southwest, the mines were placed on backgrounds ranging from loose sand to mixed sand and vegetation. The COMPASS and AHI sensors were both placed on the Twin Otter aircraft, and data was collected with the airplane as low as 700 ft and as high as 4000 ft. In this paper, the data collected on surface mines will be reviewed, and specific examples from each background type presented. Spectral detection algorithms will be applied to the data and the results of the algorithm processing will be presented.

This paper describes data collection and test results from an airborne ground penetrating radar (GPR) sensor operating as a synthetic aperture radar (SAR). Tests were undertaken to investigate the sensor's capability to support wide-area airborne minefield detection. The sensor was installed on a rotorcraft unmanned aerial vehicle (UAV). Flight tests occurred in 2002/3 at several US Army test sites containing minefields comprised of diverse types of anti-tank landmines, both metallic and low-metallic, that were buried and surface-laid. Data was collected using two side-look SAR modes: strip-map and spotlight. Strip-map mode data was collected using linear flight paths. Spotlight mode data was collected over a path surrounding the survey region allowing the sensor to collect minefield data over a full 360° view in azimuth. Data collected in strip-map mode was processed to form two-dimensional SAR imagery of the minefields. Three dimensional images were generated by processing the 360° spotlight mode data. The images were generated in a geo-referenced coordinate system to allow direct comparison of the imagery with surveyed ground truth. The sensor system is described and the flight tests that were undertaken are explained. Examples of SAR imagery from the flight tests are presented and compared to surveyed ground truth.

In support of the U.S. Army Night Vision And Electronic Sensors Directorate (NVESD), the U.S. Army Research Laboratory (ARL) has developed infrastructures, tools, and algorithms to evaluate the data set. This paper focuses on the signal processing and image formation using data from a low-frequency ultra-wideband sensor. We examine various issues that are associated with this class of SAR databases such as radio frequency interference (RFI), the effects of spectral notches, and errors in motion measurement to image quality. We show the pre-processing steps such as frequency and phase calibration, radio frequency interference extraction. We also show the application of digital spotlight technique to correct motion errors introduced by the measurement system. Finally, we show the resulting SAR imagery of various minefields.

The Army Research Laboratory (ARL) has recently examined single-polarity, synthetic aperture radar (SAR) data collected in spotlight mode at X band as part of an effort to identify land mines in radar imagery. This data set consists of several single-polarization, extremely high-resolution, spotlight-mode SAR images from multiple passes over a common target area that includes various reference reflectors as well as the landmines. In earlier investigations at Ku band, we observed that a multi-look averaging scheme could enhance the contrast between mines and background clutter. In the most recent investigation, we hypothesize that a similar behavior would be present in the X band imagery, and we demonstrate how the enhanced contrast from multi-look processing leads to improved prescreener performance under certain conditions. Results are presented in the form of receiver operating characteristic (ROC) curves for different several different prescreener parameter settings.

The Army Research Laboratory (ARL) has examined single-polarity, synthetic aperture radar (SAR) data collected in spotlight mode at X band as part of an effort to identify land mines in radar imagery. This data set consisted of several single-polarization, extremely high-resolution, spotlight-mode SAR images from multiple passes over a common target area that included various reference reflectors as well as the landmines. In particular, certain mines were placed in the scene for only a portion of the data collection, creating an opportunity for the investigation of various change detection algorithms. We describe a sub-optimal, yet effective, change detection algorithm based on a Mahalanobis-like distance metric, and we apply it to a recently collected SAR data set.

Current mine detection research indicates that no single sensor or single look from a sensor will detect mines/minefields in a real-time manner at a performance level suitable for a forward maneuver unit. Hence, the integrated development of detectors and fusion algorithms are of primary importance. A problem in this development process has been the evaluation of these algorithms with relatively small data sets, leading to anecdotal and frequently over trained results. These anecdotal results are often unreliable and conflicting among various sensors and algorithms. Consequently, the physical phenomena that ought to be exploited and the performance benefits of this exploitation are often ambiguous. The Army RDECOM CERDEC Night Vision Laboratory and Electron Sensors Directorate has collected large amounts of multisensor data such that statistically significant evaluations of detection and fusion algorithms can be obtained. Even with these large data sets care must be taken in algorithm design and data processing to achieve statistically significant performance results for combined detectors and fusion algorithms. This paper discusses statistically significant detection and combined multilook fusion results for the Ellipse Detector (ED) and the Piecewise Level Fusion Algorithm (PLFA). These statistically significant performance results are characterized by ROC curves that have been obtained through processing this multilook data for the high resolution SAR data of the Veridian X-Band radar. We discuss the implications of these results on mine detection and the importance of statistical significance, sample size, ground truth, and algorithm design in performance evaluation.

The U.S. Army Research Laboratory (ARL) has recently adapted tools and algorithms - developed as part of mission efforts - to support the U.S. Army Night Vision And Electronic Sensors Directorate (NVESD) in its attempt to determine the capability of various sensors to detect small metal and plastic mines in various clutter environments. Images from different sensors and detection results from various processing algorithms can be imported into the ARL evaluation environment for quick visualization, analysis, and performance evaluation. This paper describes the capability of this “environment” and describes how it enables us to quickly assess mine detection performance using data obtained from various sensors. The paper also discusses recent findings from these data collections and the impact that may have on detection performance.

Airborne automated target detection (ATD) and fusion experiments are frequently limited by the quality, quantity, and rapid availability of geo-registered multi-sensor, multi-platform imagery. This is especially true when working with mine targets that are smaller than the inertial measurement errors on airborne platforms. Working under the sponsorship of NVESD, we have developed and demonstrated an automated approach to inertially geo-register and ground truth imagery from multiple sensor modalities at accuracies on the order of an antitank mine dimension. Data types include ground penetrating, X-band, and Ku-band synthetic aperture radar, visible to near infrared (VIS/NIR) and longwave infrared (LWIR). This database is being used to support feature and decision-level sensor and algorithm fusion studies and to extract sensor utility metrics for a wide range of operational condition subspaces. In addition, we have standardized the format of the ground-truthed imagery products for dissemination to a larger algorithm development community and for compatibility with the U.S. Army Research Laboratory’s (ARL) Automatic Target Detection Evaluation Environment (ATD EvalEnv). This environment facilitates mine detection and fusion algorithm performance assessment across sensors, algorithms, and operational conditions. In this paper, we will discuss a process for fusion studies, including the ARL infrastructure and the techniques employed to collect and prepare the inertially co-registered imagery database.

In this paper the results of our high spatial resolution digital thermal modeling effort are compared against field data. This modeling effort created models for three cases: buried, flush-buried, and surface landmines in combination with a detailed soil model. Under a NVESD airborne countermine data collection program hyperspectral signature data were collected in the LWIR spectral region. These data formed the basis for verification of the signature results generated by our digital hyperspectral landmine model. In particular, hyperspectral thermal signature data were extracted from these airborne countermine measurement data and compared against the results of the digital modeling calculations. Spectral data were extracted from areas on and off identified mine areas and compared against similar areas from our digital model over the same spectral range. The results indicate that good correspondence with the field data was achieved. A discussion of this verification process including the field data extraction procedure, signature modeling computations, and the data comparison results will be presented in this paper.

It is practically impossible to collect an exhaustive set of minefield data for all different environment conditions, diurnal cycle, terrain conditions and minefield layouts. Such a data collection may in fact be even more expensive to ground truth, register and maintain than to acquire. This paper explores minefield synthesis using patch-based sampling of previously acquired airborne mid-wave infra-red (MWIR) images. The main idea is to synthesize a new (minefield) image by selecting appropriate small patches from the existing images and stitching them together in a consistent manner to simulate realistic imagery for different minefield scenarios. The selected patches include those from different background types, emplaced cultural clutter and different mine types. We assume a first order Markov model for the image so that the image-patch at a particular location is dependent on the characteristics of the image patch in the immediate neighborhood only. The proposed model is capable of generating any desired terrain condition (homogenous or inhomogeneous) based on a given terrain map. In addition, it supports generating different minefield layouts such as patterned or scattered minefields using mine patches from appropriate backgrounds. The paper presents representative synthesized minefield imagery and image sequences using previously collected real airborne data. Minefield image data synthesized using this procedure should be valuable in an airborne minefield detection program for evaluating most mine detection as well as minefield detection algorithms.

The aim of an anomaly detector is to locate spatial target locations that show significantly different spectral/spatial characteristics as compared to the background. Typical anomaly detectors can achieve a high probability of detection, however at the cost of significantly high false alarm rates. For successful minefield detection there is a need for a further processing step to identify mine-like targets and/or reject non-mine targets in order to improve the mine detection to false alarm ratio. In this paper, we discuss a number of false alarm mitigation (FAM) modalities for MWIR imagery. In particular, we investigate measures based on circularity, gray scale shape profile and reflection symmetry. The performance of these modalities is evaluated for false alarm mitigation using real airborne MWIR data at different times of the day and for different spectral bands. We also motivate a feature based clustering and discrimination scheme based on these modalities to classify similar targets. While false alarm mitigation is primarily used to reject non-mine like targets, feature based clustering can be used to select similar-looking mine-like targets. Minefield detection can subsequently proceed on each localized cluster of similar looking targets.

One of the primary lessons learned from airborne mid-wave infrared (MWIR) based mine and minefield detection research and development over the last few years has been the fact that no single algorithm or static detection architecture is able to meet mine and minefield detection performance specifications. This is true not only because of the highly varied environmental and operational conditions under which an airborne sensor is expected to perform but also due to the highly data dependent nature of sensors and algorithms employed for detection. Attempts to make the algorithms themselves more robust to varying operating conditions have only been partially successful. In this paper, we present a knowledge-based architecture to tackle this challenging problem. The detailed algorithm architecture is discussed for such a mine/minefield detection system, with a description of each functional block and data interface. This dynamic and knowledge-driven architecture will provide more robust mine and minefield detection for a highly multi-modal operating environment. The acquisition of the knowledge for this system is predominantly data driven, incorporating not only the analysis of historical airborne mine and minefield imagery data collection, but also other “all source data” that may be available such as terrain information and time of day. This “all source data” is extremely important and embodies causal information that drives the detection performance. This information is not being used by current detection architectures. Data analysis for knowledge acquisition will facilitate better understanding of the factors that affect the detection performance and will provide insight into areas for improvement for both sensors and algorithms. Important aspects of this knowledge-based architecture, its motivations and the potential gains from its implementation are discussed, and some preliminary results are presented.

Recently, blind tests of several automated detection algorithms operating on the NIITEK ground penetrating radar data (GPR) have resulted in quite promising performance results. Anecdotally, human observers have also shown notable skill in detecting landmines and rejecting false alarms in this same data; however, the basis of human performance has not been studied in depth. In this study, human observers are recruited from the undergraduate and graduate student population at Duke University and are trained to visually detect landmines in the NIITEK GPR data. Subjects are then presented with GPR responses associated with blanks, clutter items (including emplaced clutter), and landmines in a blind test scenario. Subjects are asked to make the decision as to whether they are viewing a landmine response or a false alarm, and their performance is scored. A variety of landmines, measured at several test sites, are presented to determine the relative difficulty in detecting each mine type. Subject performance is compared to the performance of two automated algorithms already under development for the NIITEK radar system: LMS and FROSAW. In addition, subjects are given a subset of features for each alarm from which they may indicate the reason behind their decision. These last data may provide a basis for the design of an automated algorithm that takes advantage of the most useful of the observed features.

Due to the sequential nature of software implementations of the least mean square (LMS) algorithm for processing ground penetrating radar (GPR) signals for landmine detection on uni-processor computers, search area coverage rates are lower than operational needs demand. Since hardware implementations can achieve concurrent execution through parallelization of computational elements, the penalty on execution speed due to sequential execution can be ameliorated. This paper describes a hardware implementation of LMS in a field programmable gate array (FPGA) based on a fully parallel and regular computing architecture. The architecture presented is designed specifically for data collected utilizing the NIITEK Ground Penetrating Radar (GPR), however, due to the regular architecture, this design can be easily modified to suit other sensor data formats. This paper also demonstrates quantitatively the increase in throughput achieved by an implementation on a reconfigurable FPGA compared to the same implementation in sequential software. Finally, we address how reconfigurable hardware can be integrated into an existing detection system. Here, computationally intensive tasks are scheduled to execute on hardware, thereby freeing up a processor to perform other scheduled tasks. This is achieved by embedding field programmable gate array (FPGA) board specific host-to-FPGA application programmer interface (API) calls into an existing software detection system.

In this paper we present a multi-stage algorithm for target/clutter discrimination and target identification using the Niitek/Wichmann ground penetrating radar (GPR). To identify small subsets of GPR data for feature-processing, a pre-screening algorithm based on the 2-D lattice least mean squares (LMS) algorithm is used to flag locations of interest. Features of the measured GPR data at these flagged locations are then generated and pattern recognition techniques are used to identify targets using these feature sets. It has been observed that trained human subjects are often quite successful at discriminating targets from clutter. Some features are designed to take advantage of the visual aberrations that a human observer might use. Other features based on a variety of image and signal processing techniques are also considered. Results presented indicate improvements for feature-based processors over pre-screener algorithms.

This paper evaluates the performance of a holographic neural network in comparison with a conventional feedforward backpropagation neural network for the classification of landmine targets in ground penetrating radar images. The data used in the study was acquired from four different test sites using the landmine detection system developed by General Dynamics Canada Ltd., in collaboration with the Defense Research and Development Canada, Suffield. A set of seven features extracted for each detected alarm is used as stimulus inputs for the networks. The recall responses of the networks are then evaluated against the ground truth to declare true or false detections. The area computed under the receiver operating characteristic curve is used for comparative purposes. With a large dataset comprising of data from multiple sites, both the holographic and conventional networks showed comparable trends in recall accuracies with area values of 0.88 and 0.87, respectively. By using independent validation datasets, the holographic network’s generalization performance was observed to be better (mean area = 0.86) as compared to the conventional network (mean area = 0.82). Despite the widely publicized theoretical advantages of the holographic technology, use of more than the required number of cortical memory elements resulted in an over-fitting phenomenon of the holographic network.

In this work we developed target recognition algorithms for landmine detection with ultra-wideband ground penetrating radar (UWB GPR). Due to non-stationarity of UWB signals their processing requires advanced techniques, namely regularized deconvolution, time-frequency or time-scale analysis. We use deconvolution to remove GPR and soil characteristics from the received signals. An efficient algorithm of deconvolution, based on a regularized Wiener inverse filter with wavelet noise level estimation, has been developed. Criteria of efficiency were stability of the signal after deconvolution, difference between the received signal and the convolved back signal, and computational speed. The novelty of the algorithm is noise level estimation with wavelet decomposition, which defines the noise level separately for any signal, independently of its statistics. The algorithm was compared with an iterative time-domain deconvolution algorithm based on regularization. For target recognition we apply singular value decomposition (SVD) to a time-frequency signal distribution. Here we compare the Wigner transform and continuous wavelet transform (CWT) for discriminant feature selection. The developed algorithms have been checked on the data acquired with a stepped-frequency GPR.

Ground-penetrating radar (GPR) is an effective subsurface imaging tool for solution to landmine detection, since it can detect both metal and nonmetal objects. But some times in complicated situation, for example, obliquely buried landmine will reduce its efficiency. To solve the problem, a stepped-frequency continuous-wave array antenna ground penetrating radar (SAR-GPR) system was developed. By the antenna array configuration, Common Middle Point (CMP) data can be acquired directly. Based on the CMP data, several seismic signal processing, including velocity spectrum and pre-stack migration, can be used. The velocity analysis technique based on the velocity spectrum was used here, because the velocity is required for migration. Pre-stack migration technique was used to image obliquely buried landmine in their true spatial location and physical shape in the object space. As a migration technique, pre-stack migration can also increase the signal-to-noise ratio of the survey and refocus the scattered signals. Pre-stack migration can be used by diffraction stacking based on the nonzero-offset travel time equation for a point scatterer. Amplitudes are summed along the nonzero-offset diffraction travel time trajectories. Flat ground surface and obliquely buried landmine were simulated in laboratory experiment. After the application of pre-stack migration to the CMP data acquired by SAR-GPR system, the reconstructed target image could be showed clearly.

Mathematical morphology is a field of knowledge and techniques involving the application of nonlinear image processing operations to perform image enhancement, feature extraction, and segmentation as well as a variety of other tasks. Morphological operations have previously been combined with neural networks to produce detectors that learn features and classification rules simultaneously. The previous networks have been demonstrated to provide the capability for detecting occluded vehicles of specific types using LADAR, SAR, Infrared, and Visible imagery. In this paper, we describe the application of morphological shared weight neural networks to detecting off-route, or “side attack”, mines. A pair of image sequences, both of the same scene, with and without a mine are presented to the system. The network then performs detection and decision-making on a per sequence basis.

2,4,6-Trinitrotoluene, commonly known as TNT, is an explosive used in military shells, bombs, landmines, grenades, demolition operations, and underwater blasting. It is produced in the United States only at military facilities. Accidental releases of TNT and residues in battle fields have contaminated groundwater, soil, and sand at numerous sites around the world. TNT exists in two physical forms at room temperature: droplets and crystals. The spectroscopic information conveyed depends on its physical form and the substrate on which it is deposited. Vibrational spectroscopy is a powerful tool that can be used to characterize TNT in its diverse forms. Crystallization of TNT from different solvents (acetonitrile, methanol, and water) was carried out to subsequently measure the vibrational spectra. The important nitroaromatic compound exhibits a series of unique characteristic bands that allow its detection and spectroscopic characterization. The spectroscopic signatures of neat TNT samples were determined with Raman Microspectroscopy and Fourier Transform Infrared (FTIR) Microscopy. The Raman spectra of neat TNT are dominated by strong bands at about 1365 and 2956 cm^-1. The intensity and even the presence of these bands are found to be remarkably dependent on TNT form and source.

Raman Spectroscopy is a well established tool for vibrational spectroscopy analysis. Interactions of explosives with different substrates can be measured by using quantitative vibrational signal shift information of scattered Raman light associated with these interactions. A vibrational spectroscopic study has been carried out on 2,4-DNT and 2,6-DNT crystals. Raman Microscopy spectrometers equipped with 514 nm and 785 nm laser excitation lines were used. The samples were recrystallized on different solvents (water, methanol and acetonitrile) and allowed to interact with soil samples. The interaction with sand and soil samples doped with the nitroaromatic compounds showed significant shifts in its peaks. The above information was used to detect DNT in soil using Raman Microscopy. These results will make possible the development of highly sensitive sensors for detection of explosives materials.

2,4,6-trinitrotoluene (TNT) is the most used explosive as main charge in landmines. There have been found contamination of soil and groundwater with munitions residues of TNT due to buried landmines. We are investigating the molecular structure, vibration behavior and the binding energy of TNT with the siloxane surface site of clay minerals in order to determine the spectroscopic signature of TNT in soil. Two different molecular symmetry structures were found with density functional theory (DFT) B3LYP method with 6-31G, 6-31G*, 6-311G, 6-311G*, and 6-311+G** basis sets from the Gaussian 98 systems of programs. Different deformations of the phenyl ring and distortions of the nitro and methyl groups with the ring were observed. In both structures, C₁ and C_s, the nitro groups in positions 2 and 6 are out of plane with the phenyl ring due to steric interaction with the methyl group while the nitro group in position 4 is planar to the phenyl ring. The difference between the two structures is the internal rotation of the methyl group and 2, 6-nitro groups. Comparison of the calculated energies of the two structures in several basis sets reveals that the lowest-energy geometry for the TNT structure corresponds to C_s symmetry with B3LYP/6-311+G**. FTIR spectra of TNT are presented and assigned assisted by B3LYP/6-311+G** result. The lowest-energy molecular structure of TNT was interacted with the basal siloxane surface of clay minerals to determine the binding energy (E_b) between them. The binding energy was obtained by optimizing the vertical distance, the rotational and inclination angles between TNT and siloxane surface using the B3LYP hybrid functional with different basis sets.

Among the many different signature compounds emitted from a landmine in the vapor phase, 2,4-dinitrotoluene (2,4-DNT) is the most common nitroaromatic compound in terms of detecting buried landmines, although it is a byproduct in the synthesis of TNT. 2,4-DNT is used as an ingredient in mining explosives and also prevalent on the soil surface but is somewhat seasonally dependent. The B3LYP hybrid functional was used to obtain the lowest-energy structure of both 2,4 and 2,6-DNT. Increasing basis sets from the 3-21G up to the 6-31++G (d, p) are used to predict structural parameters, vibrational frequencies, IR intensities and Raman activities for the explosives molecules. The calculated energies show that the 2,4-dinitrotoluene isomer is more stable than 2,6-dinitrotoluene isomer due to the lesser levels of steric effects between the nitro groups and the methyl group. The optimized structures were interacted with the siloxane site of clay minerals, using the density functional level of theory and the basis sets used to optimize the geometry of the DNT molecules. The binding energy (E_b) between the optimized molecules and the basal siloxane site surface of clay minerals was calculated at distances in a range between 2.5 to 8.5 Å.

Landmine detection is an important task for military operations and for humanitarian demining. Conventional methods for landmine detection involve measurements of physical properties. Several of these methods fail on the detection of modern mines with plastic enclosures. Methods based on the detection signature explosives chemicals such as TNT and DNT are specific to landmines and explosive devices. However, such methods involve the measurements of the vapor trace, which can be deceiving of the actual mine location because of the complex transport phenomena that occur in the soil neighboring the buried landmine. We report on the results of the study of the explosives subject to similar environmental conditions as the actual mines. Soil samples containing TNT were used to study the effects of aging, temperature and moisture under controlled conditions. The soil used in the investigation was Ottawa sand. A JEOL GCMate II gas chromatograph ñ mass spectrometer coupled to a Tunable Electron Energy Monochromator (TEEM-GC/MS) was used to develop the method of analysis of explosives under enhanced detection conditions. Simultaneously, a GC with micro cell ⁶³Ni, Electron Capture Detector (μECD) was used for analysis of TNT in sand. Both techniques were coupled with Solid-Phase Micro Extraction (SPME) methodology to collect TNT doped sand samples. The experiments were done in both, headspace and immersion modes of SPME for sampling of explosives. In the headspace experiments it was possible to detect appreciable TNT vapors as early as 1 hour after of preparing the samples, even at room temperature (20 °C). In the immersion experiments, I-SPME technique allowed for the detection of concentrations as low as 0.010 mg of explosive per kilogram of soil.

We report on scanning electron microscopy and energy disperse X ray fluorescence measurements of TNT deposits on dry, wet and basic Ottawa sand particles. On clean Ottawa sand particles, TNT deposits form elongated crystals that change the morphology with time. The surfaces of the crystals acquire roughness features in one month old deposits and are no longer observed in two month deposits. On wet surfaces, fresh TNT deposits form assembles that resemble wire meshes. One month old TNT deposits on wet Ottawa sand appear to cover the particles surfaces and are no longer observed in structures that resemble the crystals observed on dry deposits. Fresh TNT deposits on Ottawa sand pre treated with sodium hydroxide appear amorphous. The deposits appear to cover the particle surfaces after a month and break into thin fibers in two month old deposits.

Science & Technology International (STI) has developed, under contract with the Office of Naval Research, a system of multispectral airborne sensors and processing algorithms capable of detecting mine-like objects in the surf zone. STI has used this system to detect mine-like objects in a littoral environment as part of blind tests at Kaneohe Marine Corps Base Hawaii, and Panama City, Florida. The airborne and ground subsystems are described. The detection algorithm is graphically illustrated. We report on the performance of the system configured to operate without a human in the loop. A subsurface (underwater bottom proud mine in the surf zone and moored mine in shallow water) mine detection capability is demonstrated in the surf zone, and in shallow water with wave spillage and foam. Our analysis demonstrates that this STI-developed multispectral airborne mine detection system provides a technical foundation for a viable mine counter-measures system for use prior to an amphibious assault.

The chemical spectroscopic signature of the RDX-clay mineral complex has been investigated by means of reflectance FT-IR micro spectroscopy. The mechanical analysis method was used to separate the clay from the other soil components. The soil was obtained from the University of Puerto Rico at Mayagüez (UPRM) campus backyard. B3LYP/6-311G** calculations performed on RDX helped to determine the most stable conformations, their symmetry, and vibrational spectra. The FTIR technique confirmed the existence of two different RDX solid phases, known as the α-RDX and β-RDX, which have different symmetries and revealed significant differences in their spectra. The IR microspectroscopic study showed that the RDX-clay mineral complex and its interactions can be detected using the FTIR technique at a low concentration of 1000 part-per-millions. The results also suggest that the vibrational modes presenting changes in the different vibrational spectra correspond to the C-N and NO₂ groups. In comparison with α-RDX spectrum, the complex exhibits three bands at 740, 754 and 792 cm^-1. A 12 cm^-1 red shift is observed in this last band assign to the C-N stretching and NO₂ scissoring vibrations in the equatorial position. Differences in the spectra were also seen in the shifted bands at 942 and 953 cm^-1. These vibrational modes are assigned to the ring breathing and N-N stretching vibration in the axial position for the -phase. Comparison of the spectra of the α-RDX, the β-RDX and the RDX mixed with clay in the range from 1190 to 1700 cm^-1 clearly indicated that the FTIR technique can be used to study the interaction between RDX and clay. The results also indicate that the interaction between the RDX and the clay minerals affects mainly the NO₂ groups of the explosive molecules. It is suggested that the electron donor nitrogen atoms from RDX are interacting with the electron acceptor oxygen atoms of the siloxane surface that is present in the majority of clays.

We present the first report of a new experimental metal detector that is able to locate an underground metal object in three dimensions with an accuracy of millimeters and measure a signature to provide discrimination against frag (chaff). The ability to pinpoint the metal means that the physical excavation of the target can be conducted more quickly and safely. This detector consists of a single transmitter coil, an array of 40 receiver coils and a computer to control soil rejection and data inversion. An inversion algorithm returns the 3D location of a target with respect to the sensor head and the signature of the metal object that is largely independent of the geometry of the measurement. Tests were conducted in air, in sand and in soil using various surrogate mines and cartridge cases. Location accuracy was generally found to be very good. Several samples of a range of mine surrogates had their signatures recorded, and all samples of each type were found to have a signature falling in a very narrow band. Most of these bands are well separated, leading us to conclude that there is considerable potential for discrimination against frag. During a blind test 80% of the mines were correctly identified. We conclude that this experimental detector can accurately locate metal objects in three dimensions and provide useful information for discriminating frag from mines. This paper reports on the technology within the new detector and the early results of the performance tests conducted against surrogate mines in test lanes.

Robust, timely, and remote detection of mines and minefields is central to both tactical and humanitarian demining efforts, yet remains elusive for single-sensor systems. Here we present an approach to jointly exploit multisensor data for detection of mines from remotely sensed imagery. LWIR, MWIR, laser, multispectral, and radar sensor have been applied individually to the mine detection and each has shown promise for supporting automated detection. However, none of these sources individually provides a full solution for automated mine detection under all expected mine, background and environmental conditions. Under support from Night Vision and Electronic Sensors Directorate (NVESD) we have developed an approach that, through joint exploitation of multiple sensors, improves detection performance over that achieved from a single sensor. In this paper we describe the joint exploitation method, which is based on fundamental detection theoretic principles, demonstrate the strength of the approach on imagery from minefields, and discuss extensions of the method to additional sensing modalities. The approach uses pre-threshold anomaly detector outputs to formulate accurate models for marginal and joint statistics across multiple detection or sensor features. This joint decision space is modeled and decision boundaries are computed from measured statistics. Since the approach adapts the decision criteria based on the measured statistics and no prior target training information is used, it provides a robust multi-algorithm or multisensor detection statistic. Results from the joint exploitation processing using two different imaging sensors over surface mines acquired by NVESD will be presented to illustrate the process. The potential of the approach to incorporate additional sensor sources, such as radar, multispectral and hyperspectral imagery is also illustrated.

Laser illumination techniques continue to be investigated to support airborne detection of minefields. Polarization characteristics of manmade objects and natural backgrounds are potential discriminants for detection of mines and minefields or supplementing other detection features. NVESD sponsored development of a dual-band system prototype to investigate airborne minefield detection and determine the feasibility of packaging a system for use aboard a tactical UAV in day and night operations. The prototype contains both an infrared sensor and a laser subsystem. The laser subsystem is an active imaging near IR (NIR) assembly consisting of a laser illuminator and a receiver. The illuminator integrates multiple stacked laser diode arrays with a laser controller capable of generating short, high-energy pulses of selectable amplitude to control optical power output. The receiver includes a linear polarization analyzer and a range-gated, image intensified CCD camera. This paper presents a technical description of the active NIR subsystem, including laboratory and field testing, system modeling, and preliminary results of analysis and signal processing efforts. We describe measured reflectivity and polarization characteristics of various object classes, including mines, backgrounds, and clutter. We indicate the observed class separability due to the polarization characteristics of the object classes.

In this problem an identification vehicle must re-acquire a fixed set of suspected targets and determine whether each suspected target is a mine or a false alarm. If a target is determined to be a mine, the identification vehicle must neutralize it by either delivering one of a limited number of on-board bombs or by assigning the neutralization task to one of a limited number of single-shot suicide vehicles. The identification vehicle has the option to reload. The singleshot suicide vehicles, however, cannot be replenished. We have developed an optimal path planning and reload strategy for this identify and destroy mission that takes into account the probabilities that suspected targets are mines, the costs to move between targets, the costs to return to and from the reload point, and the cost to reload. The mission is modeled as a discrete multi-dimensional Markov process. At each target position the vehicle decides based on the known costs, probabilities, the number of bombs on board (r), and the number of remaining one-shot vehicles (s) whether to move directly on to the next target or to reload before continuing and whether to destroy any mine with an on-board bomb or a one-shot suicide vehicle. The approach recursively calculates the minimum expected overall cost conditioned on all possible values r and s. The recursion is similar to dynamic programming in that it starts at the last suspected target location and works its way backwards to the starting point. The approach also uses a suboptimal traveling salesman strategy to search over candidate deployment locations to calculate the best initial deployment point where the reloads will take place.

This paper analyzes the performance of a fast, low complexity, integer-to-integer compression scheme that is designed to give greater importance to small targets. Practical real-time operation of unmanned aerial vehicle mine/minefield detection systems has two difficult constraints. One is limited data-link bandwidth and the other is limited on-board processing power. Standard compression techniques are usually complex and tend to remove small objects from the imagery. In the imagery used for airborne mine/minefield detection, the targets are small, usually on the order of a few pixels. The region-of-interest (ROI) Wavelet Difference Reduction (WDR) compression scheme satisfies both of these con-straints and is shown to preserve detection rates of small targets. Results are compared for block-based (BB)-WDR compressed and ROI-WDR compressed and uncompressed images. The ROI -WDR process is shown to be superior to other compression conditions.

Quantum Magnetics has developed a Quadrupole Resonance (QR) system for the detection of anti-tank and anti-vehicle landmines. The QR confirmation sensor (QRCS) is a part of the Army GSTAMIDS Block 1 program and is designed to confirm the presence of landmines initially flagged by a primary sensor system. The ultimate goal is to significantly reduce the number of sites that require neutralization or other time consuming investigation into the presence of a landmine. Government tests in 2002 and 2003 demonstrated the performance of the system in a wide variety of conditions including high radio frequency interference (RFI) and piezo electric ringing (PER) environments. Field test results are presented along with an overall description of the system design and methods used to solve prior issues with RFI and PER.