Laser & Optoelectronics Progress, Volume. 56, Issue 6, 062803(2019)

Method for Extraction of Point Cloud Feature Line

Qiancheng Ren*
Author Affiliations
  • China Railway Siyuan Survey and Design Group CO., LTD., Wuhan, Hubei 430063, China
  • show less
    Figures & Tables(11)
    Flow chart of feature line detection
    Normal vector according to adjacent points
    Azimuthal angle of normal vectors on projection plane. (a) Point at boundary; (b) point in plane
    Maximum difference of azimuthal angles of adjacent normal vectors. (a) Under different range resolution; (b) under different neighborhood points
    Triangles composed by detection point and nearest neighbor points
    Cluster results of fold point detection. (a) Cluster of point normal vectors in plane; (b) cluster of point normal vectors on fold; (c) cluster of point normal vectors on curve
    Extraction of boundary. (a) Original point cloud; (b) improved RANSAC method; (c) proposed method
    Fold lines extracted by different methods under condition of plane construction. (a) Original point cloud; (b) based on curvature method; (c) based on plane method; (d) based on proposed method
    Fold lines extracted by different methods under condition of non-plane construction. (a) Original point cloud; (b) based on curvature method; (c) based on plane method; (d) based on proposed method
    Point cloud of bridge
    Extracted feature lines of bridge. (a) Feature lines extracted from first phase point cloud; (b) feature lines extracted from second phase point cloud
    Tools

    Get Citation

    Copy Citation Text

    Qiancheng Ren. Method for Extraction of Point Cloud Feature Line[J]. Laser & Optoelectronics Progress, 2019, 56(6): 062803

    Download Citation

    EndNote(RIS)BibTexPlain Text
    Save article for my favorites
    Paper Information

    Category: Remote Sensing and Sensors

    Received: Aug. 2, 2018

    Accepted: Oct. 16, 2018

    Published Online: Jul. 30, 2019

    The Author Email: Ren Qiancheng (614991092@qq.com)

    DOI:10.3788/LOP56.062803

    Topics