The identification of ground targets using synthetic aperture radar (SAR) images, in the presence of ground clutter is extremely challenging. The Moving and Stationary Target Acquisition and Recognition (MSTAR) program was an Automatic Target Recognition (ATR) System that achieved good probability of identification (Pid) success, but the false alarm rate caused by ground clutter still requires significant improvement. The goal of this proposal is to develop a comprehensive theoretical model for predicting the radar return from ground clutter based on statistical models. Parameters in those models will be either calculated by electromagnetic (EM) scattering theory or extracted from measured data to support analysis of the clutter false alarm rate of ATR systems. Theories and results will be tested in working with colleagues in AFRL target recognition programs and using their collected clutter data. All software developed in this effort will be integrated into the INSSITE scene simulation code, which is a state-of-the-art scene generation tool developed by SAIC-Champaign, TKA's teammate in this project. Commercialization of this product will support clutter predictions for forest management, crop monitoring for yields and bio-security, and flood forecasting.
Benefits: The clutter evaluation and prediction technology development effort under this SBIR will significantly improve the clutter characterization necessary to improve ATR false alarms. Additionally this technology has multiple commercial applications in the evaluation and modeling of commercial SAR systems and data.
Keywords: Radar clutter, Automatic target recognition, Synthetic aperture radar images, Modeling and simulation, Electromagnetic (EM) scattering, Rough surface scattering, Electronic Warfare (EW), Electronic Counter-Countermeasures
