Optical Processing Techniques for Satellite Pose Determination

Optical processing techniques for three-dimensional satellite position and orientation (pose) determination, so as to allow autonomous docking, are advanced. Two dimensional optical imagery from a single camera is used. The techniques introduced include an optical correlator for frame registration, a novelty filter and tracker based system for pose estimation of spin stabilized satellites from a time sequence of images, and an optical Hough transform based system for pose estimation of three-axis stabilized satellites using a single image frame. A model of the particular satellite to be docked with exists, and features on this model are compared with measured features computed from the novelty filter/tracker system, or the Hough transform based system, to estimate pose. Test results indicate very good pose estimation accuracies with both systems for various poses and lighting conditions.