Research On Surface Reconstruction From 3D Scattered Points

As an important component of Reverse Engineering (RE), surface reconstruction is a key problem in the field of CAD modeling, medical image visualization, etc. Currently, research on surface reconstruction from 3D scattered points is a focus of RE, a difficulty in RE and a key to RE. A surface reconstruction method based on deformable model is proposed in this paper after the deep study of surface reconstruction from 3D scattered points.First, a method constructing an initial triangle mesh quickly is proposed according to 3D scattered points. The point computed in the inner convex space of the 3D scattered points is defined as spherical center. The spherical radius is the shortest distance from 3D scattered points to the spherical center. Then the original triangle mesh is constructed by triangulating the uniform points which are sampled on the spherical surface. The initial triangle mesh is completed after the original triangle mesh is subdivided and projected to spherical surface. The method presented here could allow constructing initial triangle mesh with high quality by setting corresponding parameters.Then, the initial triangle mesh is considered as a deformable model, and the adaptive shortest distance approximation method is presented after 3D scattered points divided into several data segmentation. The self-restraint method is proposed to compute the unit outward normal of each mesh vertex. After the rules of vertices approximation are established, the shortest distance from each mesh vertex to corresponding 3D scattered points is computed, and then each vertex moves in outward normal direction towards the target surface. At the same time the corresponding shortest distance changes after each vertex moves. In the process of approximation, optimization methods are applied to ensure the triangle mesh quality.Finally, a subdivision method improved on (?)3 subdivision is proposed. F-vertices are generated in triangle mesh at first, then, the largest inner angle criterion is applied to the triangle mesh. The subdivision method not only achieves the goal of subdivision, but also optimizes the triangle mesh, the effect is well. The target surface is generated through the projection from mesh vertices to 3Dscattered points following the subdivision.Some instances are executed and the results show that the arithmetic in the paper is of feasible and valid.