Proc. SPIE. 8048, Algorithms and Technologies for Multispectral, Hyperspectral, and Ultraspectral Imagery XVII
KEYWORDS: Target detection, Signal to noise ratio, Hyperspectral imaging, Detection and tracking algorithms, Sensors, Reflectivity, Transform theory, Image classification, Heads up displays, Signal detection
This paper describes a novel approach for the detection and classification of man-made objects using discriminating
features derived from higher-order spectra (HOS), defined in terms of higher-order moments of hyperspectral-signals.
Many existing hyperspectral analysis techniques are based on linearity assumptions. However, recent research suggests
that significant nonlinearity arises due to multipath scatter, as well as spatially varying atmospheric water vapor
concentrations. Higher-order spectra characterize subtle complex nonlinear dependencies in spectral phenomenology of
objects in hyperspectral data and are insensitive to additive Gaussian noise. By exploiting these HOS properties, we have
devised a robust method for classifying man-made objects from hyerspectral signatures despite the presence of strong
background noise, confusers with spectrally similar signatures and variable signal-to-noise ratios. We tested
classification performance hyperspectral imagery collected from several different sensor platforms and compared our
algorithm with conventional classifiers based on linear models. Our experimental results demonstrate that our HOS
algorithm produces significant reductions in false alarms. Furthermore, when HOS-based features were combined with
standard features derived from spectral properties, the overall classification accuracy is substantially improved.
We report a new familiy of polarimetric imaging cameras based on tunable liquid crystal components. Our camera designs use a dual frequency liquid crystal tunable filter that rotates the polarization of incoming light, in front of a single linear polarizer. The unique features of this approach include fast switching speed, high transmission throughput, no mechanical moving parts, broad bandwidth, high contrast ratio, wide viewing angle, and compact/monolithic architecture. This paper discusses these tunable liquid crystal polarimetric imaging camera architectures (time division, amplitude division), the benefits of our design, the analysis of laboratory and field data, and the applicability of polarization signatures in imaging.
Many natural materials produce polarization signatures, but man-made objects, typically having more planar or smoother
surfaces, tend to produce relatively strong polarization signatures. These signatures, when used in combination with
other means, can significantly aid in the detection of man-made objects. To explore the utility of polarization signatures
for target detection applications we have developed a new type of polarimetric imaging sensor based on tunable liquid
crystal components. Current state-of-the-art polarimetric sensors employ numerous types of imaging polarimeters, the
most common of which are aperture division, micropolarizer, and rotating polarizer/analyzer. Our design uses an
electronically tunable device that rotates the polarization of incoming light followed by a single fixed oriented linear
polarizer. Its unique features include: 1) sub-millisecond response time switching speed, 2) ~75% transmission
throughput, 3) no loss of sensor resolution, 4) zero mechanical moving parts, 5) broadband (~75% of center wavelength),
6) ~100:1 contrast ratio, 7) wide acceptance angle (±10°), and 8) compact and monolithic architecture (~10 inch3). This
paper summarizes our tunable liquid crystal polarimetric imaging sensor architecture, benefits of our design, analysis of
laboratory and field data, and the applicability of polarization signatures in target detection applications.
This study describes the U.S. Army Force Protection Demining System (FPDS); a remotely-operated, multisensor platform developed for reliable detection and neutralization of both anti-tank and anti-personnel landmines. The ongoing development of the prototype multisensor detection subsystem is presented, which integrates an advanced electromagnetic pulsed-induction array and ground penetrating synthetic aperture radar array on a single standoff platform. The FPDS detection subsystem is mounted on a robotic rubber-tracked vehicle and incorporates an accurate and precise navigation/positioning module making it well suited for operation in varied and irregular terrains. Detection sensors are optimally configured to minimize interference without loss in sensitivity or performance. Mine lane test data acquired from the prototype sensors are processed to extract signal- and image-based features for automatic target recognition. Preliminary results using optimal feature and classifier selection indicate the potential of the system to achieve high probabilities of detection while minimizing false alarms. The FPDS detection software system also exploits modern multi-sensor data fusion algorithms to provide real-time detection and discrimination information to the user.