Uncooperative Spacecraft Pose Estimation Based on Intensity and Range Images Fusion

In this article, a fusion strategy of the intensity image and range image is proposed for the uncooperative spacecraft pose estimation. Considering the feature points are extracted from the intensity image, the pose estimation problem can be modeled as the perspective-n-point (PnP) problem, coupled with the point cloud data in the range image. By utilizing the planar characteristic and virtual sphere technique, the point cloud data is related to two attitude Euler angles, and the PnP problem involves the 3-D unconstrained variables. It is analyzed that the two angles and the 3-D variables are uncoupled and can be, respectively, solved by the least square method and optimization method. Then, two available attitude Euler angles are utilized to decompose the three unconstrained variables for the target pose by the singular value decomposition (SVD). Finally, the simulation and experimental results are given to validate the proposed fusion method's effectiveness for uncooperative spacecraft pose estimation.