基于超平面及遗传算法优化的网格简化

基于对目前反求工程中已有方法的分析，引入超平面的概念；通过对三角网格曲率的离散化计算，经过种子点的生长以及区域合并形成超平面。在对超平面删除后留下的空洞进行重新三角化时，利用遗传算法生成具有空间形状优化的简化模型。在此过程中以一种新的编码方式使遗传算法能够适应网格优化的特殊性；通过对变异算子的修改，保证遗传进化的有效性；将交叉算子放弃，避免了新产生的三角网格和已经存在的网格重叠和相交。通过这种方法可以调节形状的优化程度和侧重度。

Mesh Simplification Based on Super-Face and Genetic Algorithm in the Reverse Engineering

Based on the analysis of current methods a new concept of the super-face was presented. The super-face is formed through the calculation of scattered point curvatures, growth of the seed point and combination of different regions. When the hole left by the deletion of super-face was re-triangulated the genetic algorithm amended for our specific problem is used. In this process, a new coding method makes the genetic algorithm suitable for specificity of triangular mesh. The amended mutation operator can guarantee the availability of evolution. In order to avoid the intersection of triangles, the cross-over operator is given up. Based on it, the different weights to control the optimization of the new triangle mesh can be adjusted.