High Quality Point-Based Modeling of Large 3D Object

In this paper, we present a novel approach that focuses on modeling large scale 3D objects interactively. The proposed approach is a two-phases procedure: generating point-sampled geometry followed by getting smooth surface of the geometry. By defining model granularity, all the unordinary triangles be classified into three sub-groups. Therefore, the new approach can sample dynamically and generate uniform point-sampled geometry. To get a smooth surface, a method of weighted plane fitting (WPF), which can compute the normal vector of every point according to local maps from differential geometry, is introduced. We demonstrate its simplicity of implementation and its efficiency, which make it an excellent solution for modeling large 3D objects. Experiments have shown exciting results caused by the approach