Multiframe Image Enhancement and Processing for Target Discrimination

The Missile Defense Agency is seeking innovative electro-optical/infrared multiframe processing algorithms to recognize and intercept the lethal object within a ballistic missile threat train. These algorithms are critical components of Project Hercules to counter off-nominal and evolving missile threats. To address this Missile Defense Agency need, Physical Optics Corporation (POC) proposes to develop new Resolution-Enhanced Dynamic Statistical Target Advanced Recognition (REDSTAR) technology. REDSTAR is a unique integration of generic digital signal processing hardware with intelligent image processing software, which includes Multiframe Image Resolution Enhancement Framework (MIREF), Multiscale Image Contrast Enhancement (MICE) and Hierarchical Target Modeling (HTM). It will enhance resolution, detail, and contrast in low-resolution images captured by missile guidance sensors, and extract critical target information in real time to guide the kill vehicle. It can extract hidden spatio-temporal information from the image sequence to generate super-resolution images for target discrimination. REDSTAR can be applied to visible, IR, LIDAR, and MSX images acquired in boost, midcourse, or terminal phase. In Phase I, POC will demonstrate the feasibility and robustness of REDSTAR by testing against targets, target events, and backgrounds. In Phase II, POC will develop REDSTAR hardware and test REDSTAR with real sensor data in a realistic environment.