Surface model reconstruction from 3D unorganized points (points of surface or volume) is of great importance in variable fields such as computer vision

images based modeling

3D reconstruction based on images

scientific computing visualization

etc.. Many approaches have proposed to resolve the problem

such as 3D Delaunay triangulation

CDT

Qull Hull. etc. What makes the problem very difficult is that the reconstruction surface is convex and the efficiency of algorithms is not high. In this paper we present a fast model reconstruction algorithm for 3D unorganized points based on Hugues Hoppe algorithm. First

input points are divided into small logical "cubes" whose size can be decided automatically from the unorganized points according to Marching Cubes. Then the tangible planes and normal vectors at each point are calculated and all of the normal vectors are orientated to the outside of surface based on WFS(Wide First Searching). Finally

the function of iso-surface of scalar field for Marching Cubes algorithm can be obtained. In addition

the algorithm improves the efficiency of Marching Cubes by looking up tables. Finally the model can be obtained by the output of Marching Cubes composed of triangular meshes. Experimental results show the high efficiency of the algorithm. And the algorithm can be applied to not only the points of the surface but also the volume data (such as 3D scanning data

MRI data).