摘要
针对点云切片的截面数据可能存在不连续的特征曲线对拟合造成误差的情况,提出对每个截面数据将不连续的特征曲线点集分割成连续的特征曲线点集的方法。鉴于不连续特征曲线点集之间、点与点之间的欧式距离远大于点云密度,而同一个连续特征曲线点集中点与点的欧式距离趋近于点云密度,所以可以利用点云密度作为阈值,将不同的连续特征曲线点集分割开。对每个连续的特征曲线点集进行曲线拟合,得到点云切片的边界。分析结果表明,该分割算法能有效地分割出点云切片中不连续的特征曲线点集,减小曲线拟合时的误差,提高了点云切片边界提取的精度。
The section data of point cloud slices may has the case that the discontinuous feature curve cause the error in fitting. In light of this, we present a method, in which the discontinuous feature curve point set of each section data is divided into continuous feature curve point sets. In view of the Euclidean distances between the discontinuous feature curve point sets and between the points are far greater than the point cloud density, while the Euclidean distance between the points within a continuous feature curve point set approaches the point cloud density, so the point cloud density can be used as a threshold to segment different continuous feature curve point sets. Then the curve fitting is applied to every continuous feature curve points set, and the boundary of the point cloud slices is got. Analysis results show that, this segmentation algorithm can effectively segment the discontinuous feature curve point set in point cloud slices, reduce the error in curve fitting, and improve the extraction accuracy of point cloud slicing boundary.
出处
《计算机应用与软件》
CSCD
北大核心
2014年第1期222-224,245,共4页
Computer Applications and Software
关键词
点云切片
曲线拟合
点云密度边界提取
Point cloud slices Curve fitting Point cloud density Boundary extracting
作者简介
杨振清，硕士生，主研领域：三维建模。
雍永磊，工程师。
