Automatic Target Recognition from Small UAS

The capability to recognize targets in radar imagery quickly and accurately is critical for supporting battlefield Intelligence, Surveillance, and Reconnaissance (ISR). Currently, Automatic Target Recognition (ATR) systems are commonly built on large platforms with specialized radars and require trained image analysts to provide target recognition. These systems are time consuming and susceptible to human error. As a result, ATR is not widely available to unit commanders. ATR can be used in many different scenarios but when rapid decision making is a priority, the use of algorithms to automatically locate and identify targets from radar imagery allows for the quick and accurate discrimination of targets. IMSAR’s multi-mode low-Size, Weight, and Power (SWaP) radar systems provide a low-cost solution capable of providing time-critical ATR. These radar sensors provide a suite of capabilities that can enhance surveillance and increase tracking effectiveness. IMSAR proposes to partnering with Atmospheric and Environmental Research Inc. (AER) to convert current ATR software to utilize a parallel processing architecture in order to provide real time processing. This would allow for the code to run on a CUDA-enabled processor on board the airborne radar.