Yuqing Zhang,Shilong Sun,Jingjing Shang,Minghan Yang
标识
DOI:10.1109/icsip55141.2022.9886254
摘要
According to the different processing steps and ideas, the point cloud registration algorithms are classified into five categories. The representative algorithms of each type are listed to summarize the proposed time, basic ideas and advantages and disadvantages of the algorithms. For the classical algorithm of point cloud registration based on feature extraction, the voxel grid filtering method is used to down-sampling the point cloud data before the sampling consistency initial alignment algorithm. We make an experimental comparison with classical algorithms to make an analysis from subjective evaluation and objective indicators. Finally, summarize the future challenges of point cloud registration.