Algorithm Development for WFOV Mission Data Processing

ABSTRACT: Aret proposes a processing chain for FWOV OPIR data, focused on mitigation of jitter and clutter and detection of stationary and moving targets. The jitter suppression utilizes an innovative algorithm for online robust background estimation which has been successfully demonstrated on OPIR data. Tracking is done using Arets proven Bayesian track-before-detect methodology, currently in operational use in a range of Defense and Intelligence applications. These algorithms are well-understood and known to be parallelizable. The proposed effort will produce an initial parallel implementation for GPU architecture in order to predict real-time computational performance in Phase II.; BENEFIT: The proposed algorithm suite leverages previous Government investments in detection, tracking, data conditioning, and pre-processing techniques for OPIR data to rapidly achieve the desired capability at low risk. In Phase I and Phase II, the advanced algorithms proposed here will be demonstrated in increasingly realistic settings. This validation will pave the way for integration into existing persistent infrared systems (e.g. NASICs ATEP II program) and future systems such as the AFRL / SMC WFOV payload and the next generation operational OPIR system. Other applications include image stabilization and tracking small objects in wide area motion imagery (WAMI) such as from unmanned aerial vehicles in defense and surveillance industries.