摘要
提出了一种基于法矢空间离散扩张的三维点云数据特征分离算法。将三维空间中的点云数据投影到法矢空间中,通过法矢估算与离散扩张的方法从复杂模型中提取具有几何特征的曲面并将其相互分离。实验证明算法能迅速地从海量点云数据中识别并分离具有几何特征的数据点,得到单一特征曲面,并且具有较好的健壮性和算法效率。
This paper presents a new method for 3-D shape division based on k-nearest neighbor extraction.The method first project data points of 3-D space to Normal vector space, then estimate the curvature and divide point clouds by a discrete extension method which contain geometric significance and get the separated analytic surface.Experiments show that the proposed method is robust, it can extract the analytic surface from 3-D cloud data with great efficiency.
出处
《软件导刊》
2009年第12期7-9,共3页
Software Guide
关键词
点云
k邻域
分割
RANSAC
Cloud Points
K-nearest Neighbor
Space Division
RANSAC （Random Sample Consensus）
