Acta Optica Sinica, Volume. 43, Issue 12, 1228006(2023)

A Cross-Source Image Point Cloud Registration Method Combined with Graph Theory

Guanghan Chu, Dazhao Fan*, Yang Dong, Song Ji, and Zhixin Li
Author Affiliations
  • Institute of Geospatial Information, PLA Strategic Support Force Information Engineering University, Zhengzhou 450000, Henan, China
  • show less
    Figures & Tables(8)
    Process flow of proposed point cloud registration method
    Construction of kernel triangle. (a) Kernel triangles of source point cloud; (b) kernel triangles of target point cloud
    Node correspondence between two graphs in three experiments. (a) Experiment 1; (b) experiment 2; (c) experiment 3
    Comparison between satellite and aerial point cloud before and after registration
    Visualization of point cloud registration results of different algorithms
    • Table 1. Information of experimental data

      View table

      Table 1. Information of experimental data

      ExperimentPlatformImage numberImage resolutionPoint numberPoint cloud resolution /mPoint cloud densityScale differenceHoles and noise
      1Satellite20.6 m/0.8 m29518300.6082.6939Some
      Close-range3452. 8 cm196538820.281148.5
      2Satellite20.6 m/0.8 m25304550.6081.4260Many
      Close-range1933.0 cm65687560.301023.0
      3Satellite20.6 m/0.8 m14988420.6053.8134Plenty
      Close-range705.0 cm18508390.50128.8
    • Table 2. Accuracy comparison of coarse and fine registration

      View table

      Table 2. Accuracy comparison of coarse and fine registration

      RegistrationExperiment 1Experiment 2Experiment 3
      RMSEMERMSEMERMSEME
      Coarse registration7.566.937.456.6623.7921.61
      Fine registration5.165.116.395.699.028.66
    • Table 3. Comparison of registration performance between proposed algorithm and other algorithms

      View table

      Table 3. Comparison of registration performance between proposed algorithm and other algorithms

      AlgorithmExperiment 1Experiment 2Experiment 3
      Time /sChamfer distanceTime /sChamfer distanceTime /sChamfer distance
      CPD2910.920.2381153.010.235309.020.234
      RANSAC13113.220.1944331.510.1051943.100.175
      IGSP343.990.187803.090.211746.090.238
      ICP32.840.11328.630.09422.590.102
      Proposed71.040.10645.630.08037.820.094
    Tools

    Get Citation

    Copy Citation Text

    Guanghan Chu, Dazhao Fan, Yang Dong, Song Ji, Zhixin Li. A Cross-Source Image Point Cloud Registration Method Combined with Graph Theory[J]. Acta Optica Sinica, 2023, 43(12): 1228006

    Download Citation

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

    Category: Remote Sensing and Sensors

    Received: Sep. 13, 2022

    Accepted: Nov. 7, 2022

    Published Online: Jun. 20, 2023

    The Author Email: Fan Dazhao (fdzcehui@163.com)

    DOI:10.3788/AOS221702

    Topics