Algorithms for Hyperspectral Sensor Fusion of Transient Events

Persistent hyperspectral imaging has a number of important military applications including: 1) Rapid detection and identification of transient battlefield events (e.g., weapon launches); 2) Detection of low-observable moving targets not readily found using conventional spatial or spectral discriminants, or that move too quickly to be acquired and followed by low revisit rate HSI sensors; 3) Dynamic tracking and identification of gaseous effluent plumes associated with chemical weapons. To realize the full potential of the emerging generation of staring "video-HSI" (VHSI) sensors, it is necessary to develop data processing concepts that take full advantage of the novel data products these instruments provide. Space Computer Corporation proposes to extend techniques that have previously been employed to solve separate problems in HSI change detection, target tracking, chemical plume discrimination, and multisensor fusion, and to develop a new framework of algorithms and tools specifically tailored for VHSI sensors. The Phase I effort will define top-level VHSI operational concepts, develop algorithm approaches, and demonstrate algorithm utility for selected VHSI applications of military significance. In Phase II, we will implement the most promising algorithms in automated software running on COTS computer hardware, and conduct a technology demonstration with a prototype VHSI sensor system.