基于区域划分的点云全局配准研究

The research of Point cloud global registration based on region division

在分析了基于穷举思想进行全局配准的常用方法的基础上,阐明了现有方法在两点云低重叠比例情况时配准的局限性,提出了基于区域划分的超级四点全等集算法用于点云全局配准的方法,通过实验分别定性、定量地对比了本算法与四点全等集算法在不同点云数目以及不同重叠比例下的配准效果,验证了本算法在低重叠比例情况下的优越性。本算法对于几何特征不明显且重叠比例较低的点云具有较好的配准效果,对提高点云全局配准的精度及效率具有重要的意义。

Based on the analysis of the common methods of global registration based on exhaustive method,the limitations of existing methods in the case of low overlap ratio of two point cloud are clarified.A Super 4-Points congruence Sets algorithm based on region partition is proposed for the global registration of point clouds.The cost calculation is compared qualitatively and quantitatively through experiments.The registration results of the method and the 4-Points congruence sets algorithm under different number of point cloud and different overlap ratio verify the superiority of the algorithm in the case of low overlap ratio.This algorithm has a good registration effect for point cloud whose geometric features are not obvious and the overlap ratio is low.It is of great significance to improve the accuracy and efficiency of point cloud global registration.