Standoff detection of chemicals remains a crucial need for a variety of applications of importance for defense, homeland security, environmental, and industrial applications. The goal of standoff chemical sensing is to enable the identification and classification of an unknown hazardous or toxic chemical, without any operator or instrument having to come in direct contact with the chemical itself. Currently, standoff detection of chemical vapors is carried out using optical sensing techniques. Passive infrared (IR) sensors have identified chemical vapor clouds at ranges exceeding one kilometer by detecting, spectrally resolving, and analyzing scene radiance. Currently available passive IR sensors have substantial size, weight, power, and cost (SWaP-C) limitations, which reduce the number of sensors capable of being deployed in a given area, or precludes their use altogether in certain circumstances. To address these limitations, we are developing a unique passive low SWaP-C IR sensor capable of detecting chemical vapors when viewed against a cold-sky or terrestrial background. This sensor, inspired by human color vision, will use only the response through three broadband infrared optical filters to discriminate between target chemicals and background interferents. The key technology of the PBS is a commercially available pyroelectric quadrant chip sensor which contains four channels with unique bandpass IR filters installed. We demonstrate results collected using a variable temperature blackbody in the laboratory, which represents passive IR sensing against various background conditions. These results demonstrate the first step in the development of a passive bioinspired IR sensor which will use only low-cost commercially available components, and be capable of rapidly providing actionable detection of chemical vapor clouds
Passive infrared spectral sensors (7-14 um) measure brightness temperature along a line of sight, and from these
measurements the presence of a vapor cloud is deduced. How important are atmospheric temperature fluctuations due to
turbulence on the detection of vapors? We developed a stochastic simulation that uses the MODTRAN program to
explore this question. We were surprised to find that although temperature brightness fluctuations are not insignificant
compared to state-of-the-art sensor's noise (modeled as uncorrelated white noise) the effect on detection was very small
because turbulence noise is spectrally correlated and thus its effect was largely removed with a regression algorithm. In
this work we do not address the detection limit due to atmospheric interferences whose effect on detection limit may is
severe.
It is of vital interest to understand how cloud particles interact with ambient atmospheric radiation fields. We developed
a comprehensive analytical radiative transfer model for passive infrared remote sensing applicable to ground-based and
airborne sensors. We show the qualitative difference between simple non-scattering aerosols (pseudo vapor cloud) and
an aerosol cloud where scattering, absorption and emission occur. Simulations revealed two interesting observations:
aerosol cloud detection from an airborne platform may be more challenging than for a ground-based sensor, and the
detection of an aerosol cloud in emission mode is different from the detection of an aerosol cloud in absorption mode.
Performance of the matched filter and anomaly detection algorithms relies on the quality of the inverse sample
covariance matrix, which depends on sample size (number of vectors). The "RMB rule" provides the number of vectors
required to achieve a specific average performance loss of the matched filter. In this paper we extend the RMB rule to
provide the number of vectors needed to ensure a minimum performance loss (within a certain confidence). We also
review a general metric for covariance estimation accuracy based on the Wishart distribution and discuss anomaly
detector performance loss.
Hyperspectral imagery is often visualized as a three-dimensional image cube (two spatial dimensions and one spectral). When a hyperspectral sensor is set to stare at a fixed location a fourth dimension (time) is created as each new cube is sampled in time. In a ground-based stare-mode geometry each new cube has near perfect spatial registration with the previous data cubes. The problem with standard spectral-only hyperspectral detection algorithms is that they do not make effective use of temporal information. In this paper we combine temporal-differencing with temporal-spectral detection algorithms. The temporal-differencing allows for removal of most of the background prior to temporal-spectral detection. The temporal-spectral approaches combine temporal information with standard spectral-only statistical methods. By combining temporal-differencing with temporal-spectral information we are able to significantly improve detector performance and reduce the false alarm rate. We demonstrate the performance of these methods using data from the FIRST (Field-Portable Imaging Radiometric Spectrometer Technology). All the computer simulations and field data experiments show that temporal-differencing improves performance, inclusion of temporal-spectral information improves performance, and that the combination of temporal-differencing with temporal-spectral information greatly improves performance.
An algorithm using N-way analysis for the detection of multiple clouds in multi-wavelength lidar data is presented. Nway
analysis is a tool for algebraic manipulation of N-dimensional (ND) data arrays, and it allows for spatial (range),
temporal (time), and spectral (wavelength) information to be extracted simultaneously from 3D lidar data. The algorithm
tracks the spectral signal strength and location of each of the multiple clouds through time within the lidar measurements
via a method that is shown to be similar to multivariate anomaly detection. The method is data driven and can be applied
to arrays of any number of dimensions (e.g., polarization as the 4th dimension). Results of the algorithm for CO2 lidar
simulations of aerosol clouds are shown and discussed.
Knowledge of the optical properties of the biological spores in air is essential to the development of remote sensing
capabilities of biological aerosols. Recently Developed Mid-Infrared (2 to 33 μm) Variable Angle Spectroscopic
Ellipsometry has been used to evaluate the optical properties of composite bacterial Spores thin films at Air/ZnS
Interfaces. Bacterial-spores thin films was studied in attempt to determine the optical constants of biological spores and
the appropriate dielectric Effective Medium Theories that well describe these composite systems. It is shown most of
the spores lying down on the interfaces. We compared the optical constants of an individual bacterial spore calculated
from the ellipsometric data and a number of static and dynamic dielectric Effective Medium Theory models that
considering the effects of sizes, shapes, and electrical and magnetic dipole interactions of the spores. These
measurements and assessments show the validity of the models in describing the optical properties of the bacterial-spore
composite materials.
Recent experimental field trials have demonstrated the ability of both Fourier transform infrared (FTIR) and active light detection and ranging (LIDAR) sensors to detect particulate matter, including simulants for biological materials. Both systems require a reliable, validated, quantitative database of the mid infrared spectra of the targeted threat agents. While several databases are available, none are validated and traceable to primary standards for reference quality reliability. Most of the existing chemical agent databases have been developed using a bubbler or syringe-fed vapor generator, and all are fraught with errors and uncertainties as a result. In addition, no quantitative condensed phase data on the low volatility chemicals and biological agents have been reported. We are filling this data gap through the systematic measurement of gas phase chemical agent materials generated using a unique vapor-liquid equilibrium approach that allows the quantitation of the cross-sections using a mass measurement calibrated to primary, National Institutes of Standards and Technology (NIST) standards. In addition, we have developed quantitative methods for the measurement of condensed phase materials in both transmission and diffuse reflectance modes. The latter data are valuable for the development of complex index of refraction data, which is required for both system modeling and algorithm development of both FTIR and LIDAR based sensor systems. We will describe our measurement approach and progress toward compiling the first known comprehensive and validated database of both vapor and condensed phase chemical warfare agents.
A data collection experiment was performed in November of 2003 to measure aerosol signatures using multiple sensors, all operating in the long-wave infrared. The purpose of this data collection experiment was to determine whether combining passive hyperspectral and LIDAR measurements can substantially improve biological aerosol detection performance. Controlled releases of dry aerosols, including road dust, egg albumin and two strains of Bacillus Subtilis var. Niger (BG) spores were performed using the ECBC/ARTEMIS open-path aerosol test chamber located in the Edgewood Area of Aberdeen Proving Grounds, MD. The chamber provides a ~ 20' path without optical windows. Ground truth devices included 3 aerodynamic particle sizers, an optical particle size spectrometer, 6 nephelometers and a high-volume particle sampler. Two sensors were used to make measurements during the test: the AIRIS long-wave infrared imaging spectrometer and the FAL CO2 LIDAR. The AIRIS and FAL data sets were analyzed for detection performance relative to the ground truth. In this paper we present experimental results from the individual sensors as well as results from passive-active sensor fusion. The sensor performance is presented in the form of receiver operating characteristic curves.
We describe a systematic statistical analysis of the signal measured from three representative targets (grass-covered ground, woodline, and low angle sky) using a passive Fourier transform infrared spectroradiometer operating in the LWIR region (700-1350 wavenumbers/7.4-14 microns). Measurements were acquired under a wide variety of meteorological conditions including rain, snow, fog, and air temperatures. The instrumentation was operated in a temperature-controlled environment to minimize the impact of self-radiance on the measurements, and data were acquired at a variety of spectral resolutions. A standard deviation in radiance metric was developed and compared to the noise-equivalent spectral radiance (NESR) to quantify the statistical variability of the observed radiometric noise, and some assessments of the potential for background variance correction in detection algorithm development are drawn from the results.
Diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) is employed to measure the spectral properties of a chemical agent simulant, ethyl methylphosphonate (EMPA) mixed with soil. DRIFTS is a quantitative technique that provides information about the intensity of vibrational absorption bands from the analyte in terms of the Kubelka-Munk model. The spectral properties of the neat oil, the soil matrix, and mixtures of the two over a range of relative proportions will be presented and discussed vis-a-vis the surface contamination detection problem.
An experimental and modeling study performed to estimate the spectral radiance of surface contaminants is presented. The goal of the study is to address issues relevant to the passive standoff detection of surface contaminants. For this experiment, SF96 and Krylon 41325 are used as contaminant simulants and the contamination of four different surfaces (aluminum, grass, soil and plywood) is analyzed. A first order model of reflectance for surface contaminants is proposed. Measurements of spectral radiance with the CATSI system is compared with the best-fit spectra derived from the model. The experimental results agree well with the model best fits for Krylon on aluminum and grass samples. For Krylon on soil and SF96 on plywood the model best fits fail to reproduce the experimental spectra. The reasons for this discrepancy is discussed.
The HiSPEC instrument was designed to examine the potential for passive detection of sub-lethal concentrations of toxic materials and to test the potential for passive indication of biological agent in air. HiSPEC has been operating since 1999, and after substantial laboratory characterization, taken to the field several times for successful trials against known remote targets. Some subtle differences between laboratory and field performance have been diagnosed for the first time with the aid of HiSPEC's precise internal sampling system. Results of these tests may have implications for improving less sensitive passive field systems. Some recent field data is presented to indicate ultimate potential.
KEYWORDS: Clouds, Aerosols, Signal to noise ratio, Sensors, Atmospheric particles, Black bodies, Remote sensing, Infrared sensors, Information operations, Refractive index
We present a simple scaling of the SNR plots for the minimum required SNRfor detecting the emission
from an aerosol cloud. The required SNR for the detection of aerosol thermal emission is quite high, in the
order of i03 to iO' (depends on the temperature difference and the depth of the cloud) but can be achieved
with state of the art sensors equipped with large apertures and utilizing sufficient averaging.
This paper presents a multi-wavelength algorithm that utilizes the weighted mean of all possible DIAL pairs. The weights are a function of the differential absorption coefficient between the DIAL pairs. The algorithm is shown to have greater sensitivity and robustness than two- wavelength DIAL. In addition, an algorithm for estimating the column content (CL) in the presence of multiple vapors is described. The algorithm iteratively fits the lidar equation to the data by adjusting the CL of each vapor of interest. Results are shown using topographic chamber test data collected in Dugway Proving Ground in 1996 using our frequency-agile lidar sensor.
This paper extends an earlier optimal approach for frequency-agile lidar using fixed-size samples of data to include the time series aspect of data collection. The likelihood ratio test methodology for deterministic but unknown vapor concentration is replaced by a Bayesian formalism in which the path integral of vapor concentration CL evolves in time through a random walk model. The fixed- sample maximum likelihood estimates of CL derived earlier are replaced by Kalman filter estimates, and the log- likelihood ratio is generalized to a sequential test statistic written in terms of the Kalman estimates. In addition to the time series aspect, the earlier approach is generalized by (1) including the transmitted energy on a short-by-shot basis in a statistically optimum manner, (2) adding a linear slope component to the transmitter and received data models, and (3) replacing the nominal multivariate normal statistical assumption by a robust model in the Huber sensor for mitigating the effects of occasional data spikes caused by laser misfiring or EMI. The estimation and detection algorithms are compared with fixed-sample processing by the DIAL method on FAL data collected by ERDEC during vapor chamber testing at Dugway, Utah.
The effects of flight geometry, signal averaging and time- lag correlation coefficient on airborne CO2 dial lidar measurements are shown in simulations and field measurements. These factors have implications for multi- vapor measurements and also for measuring a shingle vapor with a wide absorption spectra for which one would like to make DIAL measurements at many wavelengths across the absorption spectra of the gas. Thus it is of interest to know how many wavelengths and how many groups of wavelengths can be used effectively in DIAL measurements. Our data indicate that for our lidar about 80 wavelengths can be used for DIAL measurements of a stationary vapor. The lidar signal is composed of fluctuations with three time scales: a very short time scale due to system noise which is faster than the data acquisition sampling rate of the receiver, a medium time scale due to atmospheric turbulence, and a long time scale due to slow atmospheric transmission drift from aerosol in homogeneities. The decorrelation time scale of fluctuations for airborne lidar measurements depends on the flight geometry.
Combined spectral topographical surface albedo and atmospheric transmission for a rapidly tuned CO2 airborne lidar are presented for 19 wavelengths, separated by 5 ms. The measurements were gathered over a large topographical area and show high signal-to-noise ratio (greater than 5) with single pulse acquisition for ranges up to 8 km. The temporal cross-correlation for different wavelengths is also presented and shows a relatively high correlation between wavelengths for about 20 - 50 ms. The implication of the measurements to the DIAL method is discussed.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.