Research on the self-similarity of point cloud outline for accurate compression

Point cloud has been used for industrial modeling for several years. It is created for industries purpose such as point cloud reconstruction, robot recognition, etc. Because Point Cloud Data need a lot of storage space. Especially in the application of large-scale scene reconstruction and cultural relics preservation. As Point Cloud Data compression has become an important subject. In order to process large amount of Point Cloud Data. It must guarantee a lot of point cloud data transmission in a limited bandwidth and restore the original information. The most effective way is to use the coding methods. This paper presents a novel method to compression large-scale scanning point cloud model. By calculating the feature of a point cloud model, we can find the similarity of the point cloud data and use code to replace the point which has the same feature. Qsplat method is being used to rendering and segmentation point cloud area into some similar block, and the latest RoPS algorithm is proposed to statistic the similar features of each block. By clustering block with same feature and using the block which has the similar feature to replace each-other. We can achieve the requirement of point cloud compression by remove the redundancy and recover a satisfy Point Could Data.