提出一种基于对称平面的三维点云修复方法,首先对原始点云进行重采样,并通过空间转换和密度聚类的方法获取点云的初始对称平面,然后利用镜像点云与原始点云的配准关系对初始对称平面进行优化,最终利用精细对称平面修复缺失的点云。实验结果表明本方法能够有效提取点云的对称平面,并修复缺失点云。

Aiming at the problem that the LiDAR point cloud was missing caused by occlusions, we proposed a point cloud repairing method based on the symmetric plane of point cloud. Firstly, we carried out the resample of original point data, and used space transformation and density cluster method to detect the initial symmetry plane. Then, we used the registration relationship between the original points and mirror points to optimize the initial plane. Finally, we used the mirror points obtained according to the fine symmetry plane to repair the missing part of original point. Experimental results show that the method can extract the symmetrical plane of point cloud and repair the missing LiDAR point cloud with mirror symmetrical structure.