Abstract: An imaging target tracker and a method for determining thresholds that are used to optimally distinguish a target from background. The present invention is an imaging target tracker that comprises a likelihood-based threshold selection method that uses knowledge of the location of a target to build up temporally-smoothed grey-level distribution or likelihood maps. Inner and outer histograms are formed that comprise inner and outer regions of an image, respectively, using inner, buffer and outer gates that segment the image. The inner and outer histograms are then recursively smoothed. A lookup table of actually observed pixel values is formed. A target likelihood value for actually observed pixel values is then computed. Legal start and end points of histogram segments are identified. Valid histogram segments are identified. The longest histogram segment is identified. Finally, two optimal thresholds are selected 8 and are set at respective ends of histogram segments.

Abstract: A tracking system and method that achieves system bandwidth reduction and analyzes scene information from an expanded area around a tracked target. Objects in the scene that may interfere with the target and cause degraded tracking performance are tracked to determine the level of likely perturbation of a primary tracker due to interference. The system predicts the time at which interference due to clutter objects will occur and the estimated time duration of the interference. The system uses this information to determine and variably adjust the gains of a track loop to prevent track loss while detecting and tracking maneuvers of the target.