Abstract: Based on TOF (time of flight) camera, this paper designs a pose estimation algorithm framework integrating two dimensional gray information and three dimensional point cloud information of the target to solve the relative pose between the tracking star and the non cooperative target. The algorithm framework uses the gray image and three dimensional point cloud data of the target obtained by TOF camera, and uses the arc based ellipse extraction method and ICP (iterative close point) point cloud iteration method to solve the relative pose information of non cooperative targets. This paper combines the image processing algorithm based on gray image and the pose optimization iterative algorithm based on point cloud, which improves the accuracy and robustness of the pose solution algorithm. A ground test verifies the performance of the proposed algorithm framework. The in orbit test shows that the proposed algorithm framework can stably and effectively solve the relative pose of the ultra short range space non cooperative target, provide reliable navigation information for the tracking satellite GNC subsystem, and the relative pose accuracy of TOF camera is better than 2° and 5 cm.

Key words: non cooperative target, TOF camera, pose estimation, ellipse extraction, ICP algorithm