基于特征保持和三角形优化的网格模型简化

在已有的以几何误差最小化为准则的边折叠简化算法的基础上，提出了一种新的三角网格模型简化算法．通过分析网格模型中顶点超邻域的二次误差矩阵，对模型上的重要细节特征进行定位，实现了网格简化过程中细节特征的保持．同时，在边折叠的代价函数中考虑新生成三角形的空间形状优化，并改善了简化序列的构造．算法既保持了边折叠算法快速的优点，又满足了对逼真度和网格质量的较高要求．

Mesh Simplification Based on Features Preserving and Triangles Optimization

Based on the current edge collapse algorithms for mesh simplification, in which simplification criteria are minimizing geometric error, a new algorithm was presented. In this algorithm, the sharp features on a mesh model are located through analysis of quadric error matrixes of super-neighbor of vertices on the model, so that these features can be preserved during mesh simplification. Besides, weights to control the optimization of newly created triangles during simplification are introduced into cost function for edge collapse, and the way to determine sequence of edge collapse is improved. The new algorithm not only maintains high efficiency of edge collapse algorithms but also meets the high requirements on fidelity and mesh quality.