Invariant Spectral/Spatial Techniques for Target Detection and Material Identification in Cluttered Backgrounds

In this project we will develop innovative algorithms for target detectionand tracking, material identification, and atmospheric quantification. Thealgorithms will combine spectral and spatial information in hyperspectralimages and will be applicable to the visible through short-wave infrared andthermal infrared spectral ranges. Targets will be represented using invariantphysics-based spectral subspaces that describe target signatures over spatialand temporal variability in the illumination, atmospheric, and thermalenvironments. Background clutter will be represented using adaptive spectralsubspaces that can represent arbitrary mixtures of several materials.Optimized algorithms will be derived that use the representations to detect,identify, and track low-contrast targets in unknown changing environmentswith substantial clutter. Since the algorithms are based on physical models,they can also be used to estimate parameters that describe the atmosphericconditions. The proposed algorithms are based on compact subspace representationsand can be implemented efficiently. The algorithms will be evaluated using datafrom material and atmospheric databases and imagery from hyperspectral datacollections. The new algorithms will significantly improve the capability ofsystems that process space-based electro-optical imagery.The new algorithms can be used in systems for surveillance, reconnaissance,and targeting applications. Related algorithms have important commercialapplications in areas such as biomedicine, mining, and agricultural andenvironmental monitoring.