一种基于误差控制的网格多分辨模型生成算法  

An Algorithm for Generating Multiresolution Model of Mesh Based on Error Control

作　　者：唐杰[1] 张福炎[2] 

机构地区：[1]南京大学软件新技术国家重点实验室,南京210093 [2]南京大学计算机科学与技术系,南京210093

出　　处：《计算机学报》2005年第9期1534-1540,共7页Chinese Journal of Computers

摘　　要：提出一种网格多分辨模型生成算法.该算法以最大误差L∞控制的网格简化算法为基础,通过删除边和拆分点操作进行向下和向上采样,将网格模型表示为由一个低分辨率的网格和一系列修改操作组成的多分辨模型.同其它算法相比,该算法在初始向下采样时,重点考虑了简化误差对模型精度的影响.在生成网格多分辨模型时,该算法将细化操作分解为对网格模型的几何修改信息和各细化操作之间的关系信息,确保了多分辨模型的健壮性.该算法可通过三角片数和简化误差两种方法来调整网格模型分辨率,实验结果证明了本算法的有效性.This paper puts forward a new algorithm to generate multiresolution model of mesh. The algorithm is based on a mesh simplification method controlled by L∞ error. Edge contraction and vertex split are adopted as downsampling and upsampling method respectively. The multiresolution model is represented as a base mesh and a list of refinement operations. Compared with former methods, the algorithm uses L∞ error to control the mesh simplification during the initial downsampling process. While constructing the multiresolution model, the algorithm decomposes the refinement operations into the modifications to the mesh and the dependency relations, thus ensures the robustness of the multiresolution model. Experiment results demonstrate the efficiency and stability of this method.

关 键 词：三角网格 网格简化 多分辨模型 层次细节模型 几何造型 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]