Automatic Target Recognition/Cuing Using an Unmanned Aerial Vehicle Multispectral Imaging Sensor

Using ATR/C on a UAV is a general innovation which addresses the problems of data volume, complexity and timeliness. The opportunity is to use UAVs and provide onboard ATR/C algorithms which could be implemented in the hardware as an integral part of the sensor system and be programmable to focus on specific targets, backgrounds or applications. The capability would make available, for a wide variety of problems, complex but information-intensive MSI data products for broad and general use by modestly-trained users. The Phase I objectives are to identify, characterize, catalogue, and verify (through simulation using available imagery) existing algorithms for imagery expoitation and ATR/C. A special emphasis will be on the spectral mixing analysis (SMA) algorithms developed by SETS, including the application of SMA to ATR/C using Visual-Infrared MSI; the extension of SMA to millimeter-wave, synthetic aperture radar (SAR)/radiometer, and muti-sensor data; and the investigation of incorporating spatial ATR/C algorithms with SMA algorithms.