The Research Of Surface Reconstruction Technology Based On Three-dimensional Point Cloud

With the rapid development of computer technology and modern manufacturingtechnology as well as the3D optical equipment, the obtained of high precision of3D scanningdata has been very easy. The reverse engineering has taken more and more attention in bothindustry and academic; it always became a hot spot research area in the CAD/CAMtechnology. Reverse technology is different from the traditional manufacturing technology, itcan greatly shorten the product development design cycle and copy the physical prototype in arelatively short time, and has great significance in improving the production capacity、competition and the quality of the product.Reverse engineering mainly includes data acquisition, data preprocessing and surfacereconstruction. The main work of the data acquisition is to digitize the geometry of theproduct through3D optical scanning device. The data preprocessing mainly concludes thestreamlining of data, data smoothing and the data partitioning, it has some influence to thesurface reconstruction. As an important aspect of the Reverse engineering, Surfacereconstruction aims to describe the surface of the product by a mathematical description. As awidely used technology in reverse engineering, the key step of surface reconstruction is thetriangulation. How to find the topological relationship of the point cloud and show therelationship in high quality triangular mesh is a hot research area.This paper has a deep research on the technique of the surface reconstruction based on3D pint cloud and focus on research of the direct triangulation method. Firstly, we had anintroduction of the typical triangulation algorithm. Then we introduced the K-neighborhoodspace division and the3D scattered point cloud point search algorithm. In the end of the paperwe put forward two kinds of direct triangulation method:Local optimum triangulation method based on mesh growing. By the thinking of meshthinking, we establish a spatial index to the point cloud and get the K-neighborhood of verypoint. The aim of that is to reduce the pressure of searching during the process of thetriangulation. During the process of the triangulation we add some constraints according toDelaunay triangle to make the new triangle is optimal in local.The triangular mesh reconstruction based on ball expansion is always based on the meshgrowing; the difference between them is the method of the expanding of the triangle. Wedefine a minimum bounding sphere and from the movement of the ball and the increasing of the radius to get the new triangle. This method greatly reduced the judgment of the extensionpoint, the time efficiency is greatly improved comparing to the first method.In this paper, the two methods are all verified and obtain a good result.