DeepSAR: Automating the Extraction of Tactical Information from Synthetic Aperture Radar Imagery

R-DEX Systems (R-DEX) proposes to leverage its unique expertise in artificial intelligence, machine learning, and advanced sensor solutions to develop DeepSAR, a groundbreaking software platform to automate the extraction of tactical information from Synthetic Aperture Radar (SAR) imagery for the U.S. Army and other DOD customers. As the use of SAR in the Army on manned and unmanned airborne platforms continues to grow, there is a critical need to automate the analysis of SAR imagery. Currently, human analysts are overwhelmed by the massive amounts of SAR imagery collected every day. To compound the problem, SAR imagery is harder to analyze than EO/IR imagery because the images are more susceptible to the location of scattering centers, multipath scattering, and occlusions. DeepSAR will leverage proven software to autonomously extract timely actionable information from SAR imagery. Unlike other modalities, SAR can capture images during the day and night and in all weather conditions. In addition, SAR allows detection through dense foliage, forested areas, wooded terrain, camouflage, and has applicability in multiple areas of operation. In the short term, DeepSAR will reside in a ground station; however, in the future it can be integrated directly on the airborne platform, thereby radically mitigating downlink requirements.