• Laser & Optoelectronics Progress
  • Vol. 59, Issue 12, 1228004 (2022)
Lei Geng1, Chunpeng Cao2, Zhitao Xiao1、*, and Fang Zhang1
Author Affiliations
  • 1Tianjin Key Laboratory of Optoelectronic Detection Technology and System, School of Life Sciences, TianGong University, Tianjin 300387, China
  • 2Tianjin Key Laboratory of Optoelectronic Detection Technology and System, School of Electrical and Electronic Engineering, TianGong University, Tianjin 300387, China
  • show less
    DOI: 10.3788/LOP202259.1228004 Cite this Article Set citation alerts
    Lei Geng, Chunpeng Cao, Zhitao Xiao, Fang Zhang. Multiview Point Cloud Registration Method Based on Laser Radar[J]. Laser & Optoelectronics Progress, 2022, 59(12): 1228004 Copy Citation Text show less
    Point cloud collection
    Fig. 1. Point cloud collection
    Single frame point cloud data
    Fig. 2. Single frame point cloud data
    Schematic diagram of radius filtering
    Fig. 3. Schematic diagram of radius filtering
    Pre-processing effect diagram
    Fig. 4. Pre-processing effect diagram
    Two adjacent point clouds
    Fig. 5. Two adjacent point clouds
    Block diagram of proposed method
    Fig. 6. Block diagram of proposed method
    Registration renderings. (a) Pfh-ICP; (b) Fpfh-ICP; (c) Super-4PCS; (d) NDT-ICP; (e) IMU-ICP
    Fig. 7. Registration renderings. (a) Pfh-ICP; (b) Fpfh-ICP; (c) Super-4PCS; (d) NDT-ICP; (e) IMU-ICP
    Registration effect of markers in truck
    Fig. 8. Registration effect of markers in truck
    Collection timexyzw
    10.7064510.5960950.1529890.349563
    20.6775810.6386810.2440560.270937
    30.6372510.6702530.3307270.187861
    40.5861330.6905270.4115040.101411
    50.5255220.6991040.4846540.013803
    Table 1. IMU Quaternion
    FramePointsPoints after processingPercentage reduction /%
    122390447380
    223360406783
    323765346886
    424024346986
    513433237883
    Table 2. Data preprocessing results
    MethodXYZ
    Pfh-ICP24.47149.1925.861
    Fpfh-ICP2.58928.8749.349
    Super-4PCS2.65425.5848.561
    NDT-ICP2.23448.9120.057
    IMU-ICP0.0920.0750.006
    Table 3. Average rotation error of each method
    MethodXYZ
    Pfh-ICP0.0690.0960.086
    Fpfh-ICP0.5890.8720.345
    Super-4PCS0.4350.5840.264
    NDT-ICP0.0110.0740.021
    IMU-ICP0.0030.0050.008
    Table 4. Average translation error of each method
    TimesOriginal ICPImproved ICP
    10.70256430.7015648
    20.31256240.2314568
    30.01653450.0134564
    40.00756120.0026429
    50.00265640.0023536
    60.00256440.0022203
    70.00246520.0021132
    80.0022356
    90.0022334
    100.0022267
    110.0021469
    Table 5. Sum of squared distances between original ICP algorithm and improved ICP algorithm
    Method1 /s2 /s3 /s4 /s5 /sTime consuming /s
    Pfh-ICP10.568.468.468.0735.55
    Fpfh-ICP9.617.727.277.9532.55
    Super-4PCS1.541.741.191.115.58
    NDT-ICP1.121.031.021.014.18
    IMU-ICP0.411.270.990.630.413.71
    Table 6. Point cloud continuous registration completion time
    Lei Geng, Chunpeng Cao, Zhitao Xiao, Fang Zhang. Multiview Point Cloud Registration Method Based on Laser Radar[J]. Laser & Optoelectronics Progress, 2022, 59(12): 1228004
    Download Citation