Three Dimentional Target Location From Video Images

There are strong motivations for the development of accurate, efficient algorithms for the three-dimensional location of stationary or moving targets from one or more video sensors. In this proposal we describe an effort to accomplish this task by exploiting ALPHATECH's unique expertise in all related technology areas: extraction of information from imagery, multi-object tracking, three-dimensional position estimation, GPS/INS systems, and efficient computer implementation of advanced algorithms. The proposed Target Location from Video Images system comprises the following three major parts: 1) a suite of image feature extractors chosen to provide robust performance over a wide range of sensor geometries, lighting conditions, and backgrounds; 2) state-of-the-art Multiple hypothesis Tracking algorithms to associate features that will occur in the most stressing scenarios; and 3) advanced three-dimensional position algorithms to locate the selected targets to within 2 meter accuracy by estimating the camera parameters along with the target parameter.