Surface reconstruction algorithm based on local data features

Abstract

With the development of reverse engineering device, the point cloud data, as a common and important form, is applied to the surface reconstruction domain. Since the point data is so large, scattered and unordered, the representation of point cloud and surface reconstruction is critical contents in reverse engineering system. In this paper, we redefine the representation of point cloud based on set theory, and derive the rules of neighbour relationship of data. Then, we adopt the KD-tree algorithm and improve its searching approach. In addition, the GeoSurface algorithm is proposed, which uses iterative strategy to mesh surface and uses the depth of KD-tree to control the parameters. Finally, we verify the GeoSurface algorithm. The experimental results show the GeoSurface algorithm is effective, and achieves better results in running time and quality of surface reconstruction compared with greedy and Poisson reconstruction algorithm.