Pose acquisition of non-cooperative spacecraft by point cloud template matching

This paper presents an initial pose acquisition method for non-cooperative spacecraft based on point cloud template matching. Utilizing the three-dimensional point clouds as input data, the proposed method matches the measured point cloud with an offline-generated template database to determine the coarse pose. The novelty of this study lies in the introduction of an offline template database with only one degree of freedom, resulting in a substantial reduction in both onboard storage and computational requirements. Additionally, position estimation, based on the sensor’s optical axis and target orientation bounding box features, improves ICP algorithm convergence to the global optimum compared to the traditional centroiding approach. Finally, a symmetry transformation strategy is proposed to address the pose ambiguity caused by axial symmetry. Comprehensive numerical simulations validate the effectiveness of the proposed method on sparse and uneven datasets, as well as on targets with varying shapes. Further validation using spaceborne data demonstrates the feasibility of the algorithm in practical space applications.