A Review of Point Cloud Registration Methods Based on Laser SLAM

Point cloud registration is one of the key technologies for Simultaneous Localization and Mapping of LiDAR. In the SLAM process, the LiDAR pose change leads to different poses of the point cloud data, resulting in motion distortion, and the multi-frame data point cloud is accurately registered to the same coordinate system through point cloud registration to form a completed point cloud. In this paper, the principle analysis of two different point cloud registration methods, Iterative Closest Point (ICP) algorithm and Normal Distributions Transform (NDT) algorithm, summarizes and analyzes the corresponding improved algorithms and development status, and simulates the two algorithms in MATLAB. The results show that the accuracy and robustness of the NDT algorithm are higher than those of the ICP algorithm.