基于点簇合并的多边形表面模型简化算法研究

 在计算机图形的动态实时显示技术中, 将高度密集的多边形表面模型用尽可能少的多边形逼近成为近年研究的焦点, 也出现了一批简化算法。针对存在的问题, 文中以累进网格为基础, 并将点簇合并引起的几何逼近作为研究重点, 将多种度量方法通过权因子统一到基于二次误差的模型逼近度量方法中, 并提出了改进的基于二次误差的多边形表面简化算法。实例表明该算法可快速、高质的实现逼近。该算法也适用于实时环境下的多边形网格的动态多分辨率表示。

Vertex Clustering-Based Algorithm for Simplification of Polyhedral Model

In dynamic realtime display of computer graphics, it is the critical focus for the automatic simplification of highly dense polygonal surface models to be approximate as accurately as possible by fewer polygons. Several techniques for the simplification have been developed in recent years. Aimed at the existent problems, to take the geometrical approximation error brought forth by vertex clustering into consideration and start from progressive meshes, some diverse measurement methods are unified to the quadric based measurement method of an approximation error model by means of weight factors. An improved quadric based polygonal surface simplification algorithm is put forward. Practical examples have shown it is faster and produces higher quality approximations. It also fits dynamic multiresolution in realtime render 3D engine.