STL三角形网格模型曲面特征边的提取

摘要

针对STL网格模型曲面特征边识别困难的问题,提出一种基于边和面的特征边提取方法.首先遍历网格模型,自适应地获取二面角阈值并根据该阈值识别显性特征边;然后利用最小二乘法估算曲面上顶点的平均曲率,由此计算出三角形面片的近似曲率,并利用相邻三角形的曲率差值来判别其公共的隐性特征边是否为特征边,最后采用改进的断点处特征边提取算法形成完整的特征边界.实验结果表明该方法能够有效地提取STL网格模型曲面网格的特征边,具有很好的鲁棒性.%As it is hard to recognize feature edges of the free-form surface from STL model,a method of extraction of feature edges is proposed.At first,using edge-based method,the threshold value of the dihedral angle was computed adaptively,and the explicit feature edge was identified accordingly.Secondly,using the least squares method,the approximate curvatures of triangles were obtained according to the computed average curvatures of vertexes.The feature edges were then determined in terms of the curvature difference of adjacent triangles with the implicit feature edge.Finally,a complete feature edges were generated by using an improved algorithm for extracting feature edges at the breakpoints.Experimental results demonstrate that the proposed method can effectively extract the feature edges of freeform surface from STL model,which has good robustness.