摘要
提出了一种点云数据分割方法,即在模糊数学中F聚类分析的理论基础上,利用数据点的三维坐标、法矢量和曲率构成八维特征向量,将具有类似几何特征的点集聚为一类,从而实现点云数据分割,为逆向工程中产品造型提供了一种新思路。实验证明了此方法的可行性,且方法比较简单,易于实现。
Segmentation of point clouds data is a very important problem before modeling these points on reverse engineering. Based on fuzzy cluster analysis in fuzzy mathematic,a new approach to segment point clouds data are presented. Eight dimensions feature vectors （3-dimensional coordinates,3-dimensional normal vectors and 3-dimensional curvatures） are constructed to realize segmentation of point clouds data on reverse engineering. The validity of this method is proved by the real scanned point-cloud.
出处
《轻工机械》
CAS
2007年第6期92-95,共4页
Light Industry Machinery
基金
国家自然科学基金(50475072)
关键词
逆向工程
F聚类分析
数据分割
几何特征
reverse engineering
F cluster analysis
segmentation of point clouds
geometry feature
