空间非合作目标的多视角点云配准算法研究

Research on registration algorithm of multiple-view point cloud for non-cooperative spacecraft

为了提高相邻视角间稀疏扫描点云数据配准的速度和精度,实现多视角点云精确配准,提出一种基于KD-Tree点云均匀采样简化算法,并且对传统四点算法(4-Points Congruent Sets Algorithm,4PCS)中的阈值参数进行了统一,确定了各误差阈值参数和点云密度之间的关系,通过基于姿态校正的方法有效解决了对称视角点云引起的误配准问题。仿真结果表明,该方法能够快速、有效地实现卫星稀疏点云的配准。

The point cloud data registration is one of the key technical aspects of threedimensional reconstruction for non-cooperative spacecraft, To solve the pair-wise registration issue of sparse point cloud, and align the multiple-view point cloud accurately, an optimization registration method was presented. A novel point cloud simplification algorithm using uniform sampling was proposed based on KD-Tree and the uniform relation of the threshold parameters in the 4PCS （4-Points Congruent Sets） algorithm was established via the density of the point cloud. By using the mis-registration correction method based on attitude, the mis-registration problem of the symmetry point cloud was solved. The results show that the proposed algorithm can effectively achieve good alignments of the sparse point cloud of the satellite, besides, the stability and the success rate of the algorithm is also improved.