Laser & Optoelectronics Progress, Volume. 62, Issue 14, 1428004(2025)

Improved Point Cloud Registration Method for Trajectory Tracking of Ship Unloader Grab

Zhongxuan Bai1, Yue Shen2, and Deming Kong1、*
Author Affiliations
  • 1School of Electrical Engineering, Yanshan University, Qinhuangdao 066004, Hebei , China
  • 2Hebei YSUSOFT Co., Ltd., Qinhuangdao 066000, Hebei , China
  • show less
    Figures & Tables(20)
    Real object diagram of grab bucket
    Schematic diagram of ship unloader structure and radar installation position
    Flowchart of proposed algorithm
    Visualization of feature points in grab model of ship unloader
    Calculation diagram of FPFH descriptors for grab model of ship unloader
    FPFH feature matching of grab data of ship unloader
    Schematic diagram of registration parameter optimization of FOA
    Dataset data of ship unloader grab. (a) Template; (b) target
    Bunny dataset data. (a) Template; (b) target
    Dragon dataset data. (a) Template; (b) target
    Registration performances of Bunny dataset. (a) ICP; (b) FGR-ICP; (c) RANSAC-TrICP; (d) ISSA-ICP; (e) PCR-GW; (f) proposed
    Registration performances of Dragon dataset. (a) ICP; (b) FGR-ICP; (c) RANSAC-TrICP; (d) ISSA-ICP; (e) PCR-GW; (f) proposed
    Registration performances of ship unloader grab dataset. (a) ICP; (b) FGR-ICP; (c) RANSAC-TrICP; (d) ISSA-ICP; (e) PCR-GW; (f) proposed
    Trajectory diagram of grab before and after registration
    Movement paths of grab in different directions. (a) X-direction grab path; (b) Y-direction grab path; (c) Z-direction grab path
    RMSE results of continuous registration
    • Table 1. Registration accuracy of Bunny dataset

      View table

      Table 1. Registration accuracy of Bunny dataset

      MethodBun000 and Bun045Bun000 and Bun180Bun000 and Bun270
      RMSEMAERMSEMAERMSEMAE
      ICP5.42×10-22.35×10-24.25×10-21.87×10-25.42×10-22.35×10-2
      FGR-ICP4.25×10-21.86×10-24.25×10-21.86×10-24.29×10-21.87×10-2
      RANSAC-TrICP5.40×10-32.11×10-31.13×10-24.15×10-32.23×10-29.00×10-3
      ISSA-ICP6.47×10-32.77×10-36.47×10-32.80×10-36.51×10-32.79×10-3
      PCR-GW6.50×10-32.77×10-36.47×10-32.81×10-36.51×10-32.78×10-3
      Proposed5.74×10-32.38×10-35.89×10-32.47×10-35.75×10-32.38×10-3
    • Table 2. Registration accuracy of Dragon dataset

      View table

      Table 2. Registration accuracy of Dragon dataset

      MethodDragon000 and Dragon024Dragon000 and Dragon048Dragon000 and Dragon072
      RMSEMAERMSEMAERMSEMAE
      ICP1.76×10-27.75×10-31.74×10-27.74×10-31.75×10-27.75×10-3
      FGR-ICP1.74×10-27.73×10-31.72×10-27.70×10-31.74×10-27.74×10-3
      RANSAC-TrICP5.18×10-31.93×10-35.48×10-32.04×10-38.11×10-32.97×10-3
      ISSA-ICP1.87×10-38.30×10-41.86×10-38.30×10-41.87×10-38.29×10-4
      PCR-GW1.89×10-38.39×10-41.87×10-38.08×10-41.90×10-38.19×10-4
      Proposed1.85×10-38.07×10-41.86×10-38.07×10-41.31×10-35.24×10-4
    • Table 3. Registration accuracy of ship unloader grab dataset

      View table

      Table 3. Registration accuracy of ship unloader grab dataset

      MethodRMSE /mMAE /mTime /s
      ICP0.32760.13120.16
      FGR-ICP0.20880.08463.88
      RANSAC-TrICP0.22820.09781.69
      ISSA-ICP0.03950.01610.56
      PCR-GW0.04360.01750.53
      Proposed0.03670.01450.12
    • Table 4. Performance comparison of FPFH-IFOA-ICP modules

      View table

      Table 4. Performance comparison of FPFH-IFOA-ICP modules

      MethodRMSE /mMAE /m
      FPFH0.15460.0631
      FPFH-IFOA0.05010.0181
      FPFH-IFOA-ICP0.03670.0145
    Tools

    Get Citation

    Copy Citation Text

    Zhongxuan Bai, Yue Shen, Deming Kong. Improved Point Cloud Registration Method for Trajectory Tracking of Ship Unloader Grab[J]. Laser & Optoelectronics Progress, 2025, 62(14): 1428004

    Download Citation

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

    Category: Remote Sensing and Sensors

    Received: Dec. 3, 2024

    Accepted: Feb. 7, 2025

    Published Online: Jul. 11, 2025

    The Author Email: Deming Kong (demingkong@ysu.edu.cn)

    DOI:10.3788/LOP242361

    CSTR:32186.14.LOP242361

    Topics