3D Point Cloud Mesh Reconstruction Based On Delaunay Triangulation

With the maturing of the technology of computer vision, a variety of three-dimensional reconstruction techniques are getting the attention of people. Point cloud processing and grid reconstruction technology has become a field of hot spot content. How to achieve higher precision through scanner point cloud, how to eliminate the noise points, how to make the reconstruction grid to be more perfect, to become hot spot in the research content of 3 d mesh reconstruction algorithm. This article will focus on the point cloud streamline, refactoring techniques and Openg GL simulation. Specific content as follows:In this paper, the traditional streamline algorithm based on bounding box of the point cloud and point cloud sampling streamline curvature algorithm, has been based on bounding box and an improved algorithm of streamline curvature. Firstly using the bounding box algorithm to divide the K neighborhood of point cloud, then fitting the surface of the point cloud in each little bounding box and calculate the curvature of each little box, then calculate the mean curvature H. Using the method of non-uniform grid point cloud to calculate the curvature of every little bounding box again =)n0i](i[HT K, compared ]i[HT with H, and simplify the rules. Experiments show that streamline, can realize the rapidly process point and delete redundant point.Secondly, in this paper, the growth of triangular mesh algorithm is revised, for the growth of the additional constraint of triangular mesh algorithm. The algorithm has a main idea of the traditional growth of triangular mesh reconstruction algorithm,and it was improved,additional constraints make the reconstruction of triangular mesh and it won’t produce the holes, and are close to the triangle. The results show that this algorithm processing result is perfect, can avoid the hole.Finally, summarizing the relevant works in this paper, and prospecting the future works.