Airborne LWIR hyperspectral measurements of military vehicles

Analysis of LWIR (8-13 /spl mu/m) hyperspectral data of military vehicles collected by a non-imaging sensor shows the potential of using imaging hyperspectrographs to locate and identify military vehicles. Spectral indices for soil, limestone, and vehicles are derived from in-scene data. These indices are then applied to other data sets to locate vehicles in sand and vegetation. This method consistently locates the vehicles, day or night, regardless of thermal contrast with the surroundings. Although these indices are robust, the results show that the sensor´s Signal-to-Noise Ratio (SNR) of more than 1000 is barely adequate. Higher SNR will be required to accurately locate military vehicles with a low number of false alarms. An imaging LWIR hyperspectrograph is under construction to further research this area.