中国光学, 2017, 10 (3): 348, 网络出版: 2017-06-06   

基于改进的局部表面凸性算法三维点云分割

Improved local convexity algorithm of segmentation for 3D point cloud
作者单位
1 中国科学院 长春光学精密机械与物理研究所 激光与物质相互作用国家重点实验室, 吉林 长春 130033
2 中国科学院大学, 北京 100049
摘要
点云分割是点云分类、识别以及三维重建等处理的基础, 分割结果对后续应用影响巨大。本文提出利用连通点集改进局部表面凸性算法中邻近点关系的方法, 解决目前激光三维成像系统点云分割算法在处理复杂环境散乱点云时存在分割过度及分割不充分的问题, 通过主顶点与周围点构成连通集, 作为分割判断局部子点集, 形成有效分割区域。该方法解决了常用点云分割方法无法对形状不规则物体进行有效分割的问题, 提高了分割精度。算法实验结果表明, 相比于最小切割算法和区域生长算法, 基于连通点集的改进局部表面凸性算法对实际路面环境信息的分割效果更好, 并能在一定程度上避免分割过度和分割不充分的情况, 证明该方法适用于复杂环境散乱点云数据分割。
Abstract
Segmentation for point cloud is the basis of classification, recognition and reconstruction of point cloud datasets and the segmentation result plays an important role in following research. In this paper, we propose a method using connected point sets to analyze and improve the relationship between adjacent points in the local convexity segmentation, to solve problems of oversegmentation and undersegmentation when using the existing algorithms to segment scattered point cloud data in complex environment in 3D laser imaging system. By this method we use the main vertex and neighbors to constitute connected point sets which can be local point subsets of segmentation and form the effective segmented regions. The method solves the problem of the irregular object′s segmentation, which can not be accomplished by common methods, and improves the accuracy of segmentation. Compared with the min-cut based segmentation and region growing segmentation, the improved local convexity segmentation of connected point sets is better for segmentation results of actual road information, and it can avoid oversegmentation and undersegmentation to some extent. It proved that this method is suitable for segmentation of scattered point cloud data in complex environment.

王雅男, 王挺峰, 田玉珍, 孙涛. 基于改进的局部表面凸性算法三维点云分割[J]. 中国光学, 2017, 10(3): 348. WANG Ya-nan, WANG Ting-feng, TIAN Yu-zhen, SUN Tao. Improved local convexity algorithm of segmentation for 3D point cloud[J]. Chinese Optics, 2017, 10(3): 348.

本文已被 4 篇论文引用
被引统计数据来源于中国光学期刊网
引用该论文: TXT   |   EndNote

相关论文

加载中...

关于本站 Cookie 的使用提示

中国光学期刊网使用基于 cookie 的技术来更好地为您提供各项服务,点击此处了解我们的隐私策略。 如您需继续使用本网站,请您授权我们使用本地 cookie 来保存部分信息。
全站搜索
您最值得信赖的光电行业旗舰网络服务平台!