局部中轴约束的散乱点云特征提取方法

Method for Extracting Scattered Point Cloud Features of Local Medial Axis Constraints

逆向工程领域内,为提升散乱点云特征提取结果的准确性,提出一种基于局部中轴判断采样曲面形态的方法。通过三维Voronoi剖分处理初始点云数据,基于任一给定样点邻近区域内的Voronoi极点模拟局部中轴面,随后求解极点坐标的协方差矩阵,通过主元分析、特征值平面拟合等方法求解局部中轴的形位分布特性,进而判断采样曲面在给定样点处的形貌特点,实现对棱边、尖角区域样点的识别与提取。试验结果表明,该方法适用于不同采样密度的点云,可显著提升棱边、尖角区域特征提取结果的准确性。

In order to solve the problem of scattered point cloud feature extraction, a method based on local central axis distribution to judge the shape of the surface is proposed. Firstly, the initial point cloud data is processed by three-dimensional Voronoi splitting,based on the local sampling region Voronoi pole simulation central axis, then the covariance matrix of local pole coordinates is solved, and the number of local central axes is analyzed by principal component analysis, eigenvalue plane fitting, etc. Shape, position and other characteristics, and then determine the shape characteristics of the sampling surface at a given sample point,to achieve the identification and extraction of edge and sharp comer samples. The experimental results show that the method is suitable for point clouds with different sampling densities,which can significantly improve the accuracy of the recognition results of the feature points of the curved edges and sharp comers.