Laser & Optoelectronics Progress, Volume. 58, Issue 8, 0810014(2021)

Three-Dimensional Point Cloud Registration Method with Low Overlap Rate

Yuan Zhang, Xiaoyan Li*, and Xie Han
Author Affiliations
  • School of Data Science and Technology, North University of China, Taiyuan, Shanxi 030051, China
  • show less
    Figures & Tables(11)
    Main flow chart of proposed algorithm
    Registration results of Bunny000 and Bunny045. (a) Initial position; (b) registering using method of Ref. [11]; (c) registering using method of Ref. [21]; (d) registering using proposed method
    Registration results of Dragon000 and Dragon024. (a) Initial position; (b) registering using method of Ref. [11]; (c) registering using method of Ref. [21]; (d) registering using proposed method
    Registration results of Bunny045 and Bunny090. (a) Initial position; (b) registering using method of Ref. [11]; (c) registering using method of Ref. [21]; (d) registering using proposed method
    Registration results of Dragon000 and Dragon048. (a) Initial position; (b) registering using method of Ref. [11]; (c) registering using method of Ref. [21]; (d) registering using proposed method
    Registration results of Happy048 and Happy096. (a) Initial position; (b) registering using method of Ref. [11]; (c) registering using method of Ref. [21]; (d) registering using proposed method
    Registration results of Chair1 and Chair2. (a) Initial point cloud 1; (b) initial point cloud 2; (c) registering using proposed method
    Registration results of Motobike1 and Motobike2. (a) Initial point cloud 1; (b) initial point cloud 2; (c) registering using proposed method
    • Table 1. Comparison of point cloud registration results with high overlap rate

      View table

      Table 1. Comparison of point cloud registration results with high overlap rate

      Point cloudRegistration algorithmLCPRunning time /s
      Go-ICP0.76073.00
      Super4PCS0.86131.25
      Bunny000 and Bunny045Method of Ref. [11]0.93010.05
      Method of Ref. [21]0.96412.98
      Proposed method0.97912.27
      Go-ICP0.72863.50
      Super4PCS0.80323.74
      Dragon000 and Dragon024Method of Ref. [11]0.80613.85
      Method of Ref. [21]0.87512.15
      Proposed method0.92011.80
    • Table 2. Comparison of point cloud registration results with low overlap rate

      View table

      Table 2. Comparison of point cloud registration results with low overlap rate

      Point cloudRegistration algorithmLCPRunning time /s
      Go-ICP0.59895.55
      Super4PCS0.67635.89
      Bunny045 and Bunny090Method of Ref. [11]0.72513.76
      Method of Ref. [21]0.78312.23
      Proposed method0.83412.09
      Go-ICP0.47983.43
      Super4PCS0.55932.96
      Dragon000 and Dragon048Method of Ref. [11]0.58717.49
      Method of Ref. [21]0.4938.34
      Proposed method0.75613.92
      Go-ICP0.39967.05
      Super4PCS0.41423.37
      Happy045 and Happy090Method of Ref. [11]0.45915.86
      Method of Ref. [21]0.4728.12
      Proposed method0.72810.45
    • Table 3. Registration performance of real 3D scan data

      View table

      Table 3. Registration performance of real 3D scan data

      Point cloudLCPRunning time /s
      Chair1 and Chair20.80412.49
      Motobike1 and Motobike20.77311.15
    Tools

    Get Citation

    Copy Citation Text

    Yuan Zhang, Xiaoyan Li, Xie Han. Three-Dimensional Point Cloud Registration Method with Low Overlap Rate[J]. Laser & Optoelectronics Progress, 2021, 58(8): 0810014

    Download Citation

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

    Category: Image Processing

    Received: Jul. 7, 2020

    Accepted: Sep. 15, 2020

    Published Online: Apr. 12, 2021

    The Author Email: Li Xiaoyan (2602029654@qq.com)

    DOI:10.3788/LOP202158.0810014

    Topics