Rapid Surface Reconstruction Algorithm from Dense Point Cloud

In order to triangulate a dense scattered point cloud efficiently,a novel surface reconstruction method is proposed.The algorithm presented consists of two steps:Firstly,an initial triangle mesh is constructed by repeating a simple advancing front rule.Then,initial triangles are subdivided to obtain the final accurate surface and triangle vertexes are refined by means of Multilevel B-spline fitting.Compared with other popular methods,the subdivide speed is a significant advantage of the algorithm.Besides,the algorithm guarantees that the distance from original points to the result mesh is within a predefined tolerance.Several experiments based on real scan data are used to evaluate the efficiency of this algorithm and the speed advantage is demonstrated by the comparison with other popular algorithms.