A Study On Key Technology Of Triangular Mesh Reconstruction From Unorganized Point Cloud

Three-dimensional model presents a strong perspective of the real environment, so that the surface reconstruction technique of three-dimensional model has a wide range of application, such as reverse engineering, computer aided geometric design, computer aided manufacturing, data visualization, machine vision, virtual reality, medical industry and so on. With the continuous development of the measurement equipment, getting the point cloud data of the real objects and the scenes is no longer the major factor limiting the development of the surface reconstruction technology. Because of lacking the topology of point cloud data, the key point of the surface reconstruction from unorganized point cloud is to establish the correct topology. In addition, large-scale scanning device brings massive problems about the point cloud data storage, transmission, simplification, which need new adjustment of original resurfacing technology.A large number of surface reconstruction algorithms are mainly divided into three classed, including reconstruction method based on Delaunay triangulation, based on implicit surface reconstruction and regional expansion methods. Among them, the mesh reconstruction method of Delaunay triangulation has become a hot study point for its excellent quality and other advantages. The regional expansion method based on the mesh reconstruction method of Delaunay triangulation, used in this study is faster and easier to achieve. Based on the above study background of the surface reconstruction, the study mainly completes the following tasks:(1) Three algorithms of the triangle mesh reconstruction of Delaunay triangulation are analyzed. The advantages and disadvantages of the divide and conquer method, the point by point insertion method and the growth triangle mesh method are analyzed. Point by point insertion method is faster than divide and conquer method, while divide and conquer method requires more memory space, growth triangle mesh method uses less. The Delaunay triangulation is programmed to demonstrate its characteristics.(2) This study presents an adaptive three-dimensional grid division, according to the evaluation of areal density point cloud data based on the theory of polyhedral. Based on this division method, the triangular mesh reconstruction implementation process is given by using grid as the basic unit. First, no points difference macroscopic estimation method is used to obtain three-dimensional grid side, then the basic unit of data points is selected as the seed point. The triangle side lengths is set as approximately normal 6 neighborhood constraint to built the initial triangular mesh, finally, outside expansion of the triangular mesh reconstruction is completed step by step. Experiment results show that the method proposed in this study is fast. The method in the study is for dense point cloud reconstruction, experiment examples show that when the points more than 20,000, simplify rate reaches to 93.97% in just 14.50 s. Compared with other methods, the speed of reconstruction is improved by 57.33%. In this study, it is robust for testing both the standard data and the laboratory self-test data. In addition, the method also has the advantage of mixing a simplified process and the reconstruction together.