复杂模型的视点依赖简化算法

View-Dependent Simplification Algorithm for Complex Models

为了达到在虚拟环境的漫游系统中对复杂模型进行实时绘制的目的 ,通常采用多层次细节模型来解决这一问题 .给出一种即与视点相关又无须对原来的复杂模型进行预处理的网格简化算法 .该算法采用精细 (refinem ent)法 .每次绘制前都从最粗糙的模型开始 ,根据视点位置、视方向和模型的每个部分产生的误差 ,对影响视觉效果大的部分进行细分 (subdivision) ,逐步产生可显示的模型 .实验表明 ,本文提出的网格简化算法能在对视觉效果影响不大的前提下大大降低需要绘制的三角形数量 .

In order to real-time render complex object in t he walkthrough system of virtual environment,the lever-of-detail(LOD) is of ten used as an important tool. An algorithm for mesh simplification both view-d ependent and without preprocessing for original complex model was presented in t his paper and based on the refinement method.Before rendering at each frame,it begin with coarsest model and subdivision patches which were more affluence wit h vision effect based on the viewpoint,view direction and approximation error i n each patch.Finally a displayable model was.Examples show that the number of triangle needed rendered can decrease greatly,while the vision effect is small .