• Laser & Optoelectronics Progress
  • Vol. 61, Issue 10, 1015003 (2024)
Hao Tang1、2, Dong Li1、2、*, Cheng Wang2, Sheng Nie2, Jiayin Liu1, and Ye Duan1
Author Affiliations
  • 1Faculty of Land Resources Engineering, Kunming University of Science and Technology, Kunming 650093, Yunnan , China
  • 2Key Lab of Digital Earth Science, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China
  • show less
    DOI: 10.3788/LOP232000 Cite this Article Set citation alerts
    Hao Tang, Dong Li, Cheng Wang, Sheng Nie, Jiayin Liu, Ye Duan. Global Registration Method for Laser SLAM Point Clouds Based on Graph Optimization[J]. Laser & Optoelectronics Progress, 2024, 61(10): 1015003 Copy Citation Text show less
    Loop closure edges in the pose graph
    Fig. 1. Loop closure edges in the pose graph
    Global registration flow chart
    Fig. 2. Global registration flow chart
    FAST-LIO 2 algorithm result of Data 1
    Fig. 3. FAST-LIO 2 algorithm result of Data 1
    FAST-LIO 2 algorithm result of Data 2
    Fig. 4. FAST-LIO 2 algorithm result of Data 2
    FAST-LIO 2 algorithm result of Data 3
    Fig. 5. FAST-LIO 2 algorithm result of Data 3
    FAST-LIO 2 algorithm result of Data 4
    Fig. 6. FAST-LIO 2 algorithm result of Data 4
    Comparison of point clouds before and after trajectory drift error optimization for Data 1 and Data 2
    Fig. 7. Comparison of point clouds before and after trajectory drift error optimization for Data 1 and Data 2
    Comparison of point clouds before and after trajectory drift error optimization for Data 3 and Data 4
    Fig. 8. Comparison of point clouds before and after trajectory drift error optimization for Data 3 and Data 4
    Scene comparison of Data 1 before and after optimization
    Fig. 9. Scene comparison of Data 1 before and after optimization
    Scene comparison of Data 2 before and after optimization
    Fig. 10. Scene comparison of Data 2 before and after optimization
    Scene comparison of Data 3 before and after optimization
    Fig. 11. Scene comparison of Data 3 before and after optimization
    Scene comparison of Data 4 before and after optimization
    Fig. 12. Scene comparison of Data 4 before and after optimization
    Data nameTrajectory length /mPoint cloud frame rateScanning time /s
    Data 14102018797
    Data 229302700265
    Data 335803420344
    Data 420103180322
    Table 1. Description of input data

    Data

    name

    Processing

    time /s

    Number of coarse

    optimization iterations

    Number of fine

    optimization iterations

    Data 174828
    Data 2100045
    Data 3123135
    Data 4117335
    Table 2. Description of experimental results
    Optimization stateData 1Data 2Data 3Data 4
    Before optimization2.66710.34819.0183.412
    Drift error optimization0.2310.3290.3980.223
    After coarse optimization0.1710.2840.3420.193
    After fine optimization0.1580.2110.2180.157
    Table 3. Average value of the distance RMSE of the matching point pairs
    Hao Tang, Dong Li, Cheng Wang, Sheng Nie, Jiayin Liu, Ye Duan. Global Registration Method for Laser SLAM Point Clouds Based on Graph Optimization[J]. Laser & Optoelectronics Progress, 2024, 61(10): 1015003
    Download Citation