摘要
对测量获得的三角网格模型进行分块。首先根据三角片中心点的高斯曲率和平均曲率的符号将三角片标示成 8种类型。然后任取一个三角片作为生长点 ,沿着边界向外扩展“生长”,将具有相同类型的三角片组成一个连通的块。最后采用两种方法对块实施优化 :根据平均曲率差值最小的原则 ,将小块融合到邻近块中 ,用两点间的最短路径来光滑边界。本方法中的三角网格和块的拓扑均可以是任意的 。
A data segmentation algorithm is proposed for triangular mesh models obtained from measurement equipments. Firstly, triangles are labeled for eight fundamental types by the sign of their centers of the Gussian curvature and the mean curvature. Triangles of the same type connecting with each other make up a segmented region. Then, the regions are extracted from a seed triangle and “growing” from the borders. Finally, two methods are utilized to optimize the regions. A region with little triangulation is united into the neighbor region if they have the minimum mean curvature difference. Besides, region borders are smoothed through the shortest path between two vertexes. The segmentation algorithm can be applied to arbitrary meshes and the extracted region border on feature lines of real models.
出处
《南京航空航天大学学报》
EI
CAS
CSCD
北大核心
2003年第6期653-658,共6页
Journal of Nanjing University of Aeronautics & Astronautics
基金
航空科学基金 ( 0 0 H5 2 0 69
0 1 H5 2 0 5 1 )
江苏省创新人才培养基金 ( BK2 0 0 1 40 8)资助项目
