Locally adaptive processing of sensor imagery for target acquisition and handoff

Image processing techniques, which are adaptive to local statistics, show promise for real-time improvement of image quality measures important to target acquisition and handoff. Specific processing methods are evaluated in terms of enhanced contrast, resolution and signal-to-noise ratio. The evaluation includes analytical results for simple deterministic image functions and experimental results from application of the processors to typical television and forward looking infrared imagery. Methods evaluated include local area gain control, median filtering, adaptive low-pass filtering and windowing with edge detection. A major constraint in the selection of specific processors was a requirement for real-time or near real-time implementation. Parallel and cascade combinations of multiple processing steps are evaluated.