光学学报, 2018, 38 (1): 0112001, 网络出版: 2018-08-31   

基于体素生长的点云结构直线段提取 下载: 1257次

Extraction of Structure Line Segments from Point Clouds Using Voxel-Based Region Growing
作者单位
信息工程大学导航与空天目标工程学院, 河南 郑州 450001
摘要
针对现有结构直线段提取方法存在的效率低下或准确程度不足等问题,提出了一种基于体素生长的点云结构直线段高效提取方法。首先,对点云进行体素化剖分与平面分割,并以体素为单位进行邻域判断,实现对结构直线段分布区域的筛选;然后,采用基于体素的区域生长对结构直线段的分布区域进行分割;最后,依据结构线段分布区域的范围以及其所在平面的数学方程实现其提取和优化,并进行精度评定。进行了实验测试,利用多组点云数据验证了本方法的有效性,利用对比实验验证了本方法的精度和高效性。实验结果显示:相比现有方法,该方法在效率上提高了10倍以上,在精度上提高了0.25倍左右,证明提出的方法可以准确高效地得到较为理想的点云结构直线段提取结果。
Abstract
Aim

ing at the problems remaining in existing extraction method of structure line segment such as low efficiency and reliability, an efficient extraction method for structure line segments from point clouds through voxel-based region growing is proposed. Firstly, point cloud is voxelized and segmented, and the distributing regions of structure line segments are recognized through voxel-based nearest neighbors searching. Then, the distributing regions of structure line segments are segmented through voxel-based region growing. Finally, corresponding line segments are extracted and optimized in these segmented regions according to the equations of their supporting planes, and the accuracy of the results are assessed. The test experiments are performed, and the feasibility, accuracy and efficiency of the proposed method are verified by obtaining the processing and comparison results of adopted point clouds. The experimental results indicate that the efficiency is increased more than 10 times and the accuracy is improved about 0.25 times, which verifies that the proposed method has high accuracy and efficiency as well as the ability of achieving relatively more ideal results.

李明磊, 宗文鹏, 李广云, 王力. 基于体素生长的点云结构直线段提取[J]. 光学学报, 2018, 38(1): 0112001. Minglei Li, Wenpeng Zong, Guangyun Li, Li Wang. Extraction of Structure Line Segments from Point Clouds Using Voxel-Based Region Growing[J]. Acta Optica Sinica, 2018, 38(1): 0112001.

本文已被 6 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!