Research on the Influence of Point Cloud Geometric Features on Compression Algorithm

Although point cloud has abundant information and can accurately describe objects' surface shape and structure, the amount of data has always been an essential factor restricting its wide application. In order to retain adequate information at the same time, it is necessary to reduce the amount of point cloud data as much as possible, so a variety of point cloud compression algorithms have appeared. This paper mainly studies the effect of geometric features of point clouds on compression algorithms. First, we choose four classical algorithms based on different compression strategies as the primary research objects and explain their compression principles. Then, we take the underwater multi-beam point cloud in the riprap area as the experimental object, use four algorithms to test, and compare the compression effect. Finally, combined with the experimental results, we analyze the adaptive effects of different feature attributes on the compression algorithm. The experimental results show that the uniform compression algorithm can maintain the overall features of the point cloud but will lose more details. The curvature compression algorithm can keep the details but dramatically impacts data integrity. The octree compression algorithm can maintain the whole data but will lose local details. Although the information content compression algorithm has a good compression effect, its execution efficiency is relatively low.