测绘通报 ›› 2020, Vol. 0 ›› Issue (5): 47-50,68.doi: 10.13474/j.cnki.11-2246.2020.0143

• 学术研究 • 上一篇    下一篇

融合主成分分析和局部邻域的平面点云去噪

陈建华1,2   

  1. 1. 河海大学水利水电学院, 江苏 南京 210098;
    2. 江苏河海工程技术有限公司, 江苏 南京 210098
  • 收稿日期:2019-10-24 修回日期:2020-01-14 出版日期:2020-05-25 发布日期:2020-06-02
  • 作者简介:陈建华(1982-),男,硕士,讲师,主要从事水工建筑物、桥梁和地铁等安全监测方面的研究。E-mail:cjh@hhu.edu.cn
  • 基金资助:
    国家重点研发计划(2018YFC0407104;2016YFC0401601);国家自然科学基金重点项目(51739003)

Planar point cloud denoising using the methodology of fusion of principal component analysis and local neighborhood

CHEN Jianhua1,2   

  1. 1. College of Water Conservancy and Hydropower Engineering, Hohai University, Nanjing 210098, China;
    2. Jiansu Hohai Engineering Technology Co., Ltd., Nanjing 210098, China
  • Received:2019-10-24 Revised:2020-01-14 Online:2020-05-25 Published:2020-06-02

摘要: 大型水工建筑物和城市地下轨道是国家的重要基础设施,人工变形监测耗时费力,利用激光雷达扫描对其进行自动化的变形监测已成为研究热点。而此类建筑物具有较多的平面结构特征,为了高效去除平面结构扫描点云中的噪声点,本文提出了一种融合主成分分析和局部邻域的噪声剔除方法,首先利用主成分分析剔除全局噪声,然后利用局部邻域对局部小噪声进行剔除,结合实例对该方法进行了验证。应用结果表明,该方法效率高、算法简单可行,具有较高的推广应用价值。

关键词: 变形监测, 主成分分析, 局部邻域, 去噪, 平面特征

Abstract: Large hydraulic structures are very important infrastructures of the nation. It has become a research hotspot to realize the deformation monitoring using laser scanning technique. Buildings are with many planar structure characteristics, and in order to remove the noise efficiently from the point cloud of planar structures, the methodology combining principal component analysis and local neighborhood is proposed. Firstly, the global noise is eliminated using principal component analysis, and the local noise is detected and removed by considering its local neighborhood. The methodology is verified by the measured data and the results demonstrate that the proposed methodology has the advantages of high-efficiency and feasibility, which is of high application value.

Key words: deformation monitoring, principal component analysis, local neighborhood, denoising, planar features

中图分类号: