The Research Of Surface Reconstruction Based On The Point Cloud Data Of Ground 3D Laser Scanning

The application of 3d laser scanning technology is becoming more and more abundant because of the breakthrough of 3d laser scanning technology theory and the rapid development of its corresponding hardware and software facilities.Now a new treatment model for 3d entity modeling has been developed by Ground-based 3d laser scanning technology with its unique measurement method.So the question of how to handle point-cloud data quickly and efficiently has been became the new direction of research.The question of point cloud reconstruction of 3d laser point cloud data on the ground is researched systematically in this paper.What is more,the corresponding improvement algorithms and optimization measures of this research are proposed.The main research of this paper is as follows:(1)The research situation at home and abroad of ground-based 3d laser scanning technology and point-cloud reconstruction theory are summarized in this paper.Then the basic principle,classification and structure of ground-based 3d laser scanning technology are described.In addition,the processing flow of point cloud data is introduced.(2)In this paper,an algorithm based on curvature estimation for coarse registration and accelerating ICP algorithm for precise registration is proposed through the study of the coarse and precise registration of 3d point cloud data of the ground.And then an experimental analysis combined with the sample data collected is operated by MATLAB software.At last,the advantages of the algorithm in the matching effect are presented.(3)The types of 3d point cloud date surface reconstruction algorithm are generalized and the theory of B-spline surface reconstruction is emphasized in this paper.Then the algorithm of B-spline surface reconstruction is improved based on least square method,the advantages and disadvantages of the improved algorithm are proposed combining experimental data with MATLAB.(4)The BP neural network is optimized for four aspects,based on this and moving least-squares methods,the improved BP neural network algorithm based on MLS is proposed in this paper.Then,an experiment was carried out with MATLAB software in order to verify the superiority of the improved algorithm by comparing the results of the experiment with the type B-spline surface.