Infrared and Laser Engineering, Volume. 51, Issue 12, 20220114(2022)

Robust multi-view registration method for narrow scenes

Fei Liu... Hanlin Huang, Tian Yang, Wenbo Li and Yang Yang |Show fewer author(s)
Author Affiliations
  • State Key Laboratory of Mechanical Transmissions, College of Mechanical and Vehicle Engineering, Chongqing University, Chongqing 400044, China
  • show less
    Figures & Tables(18)
    Flow chart of proposed algorithm
    Flow chart of multi-strategy registration algorithm
    Structure of pose graph
    Diagram of structured light measuring equipment. (a) High-speed projectors;(b), (c) Industrial camera;(d) Scanning head
    Tooth model data. (a) Complete model and four single-perspective point cloud; (b) Scanning path
    Result comparison of adjacent point cloud registration. (1) ICP; (2) Generalized ICP; (3) FPFH-based method; (4) MRA
    Convergence curves of four algorithms. (a) View 37 and 38; (b) View 210 and 211
    Comparison of overall results of adjacent point cloud registration. (a) ICP; (b) Generalized ICP; (c) FPFH-based method; (d) MRA
    Comparison of overall results of multi-view point cloud registration. (a) Ref.[15] method; (b) Proposed method
    Comparison of loop closure detection results. (a) View 132 and 192; (b) View 96 and 244; (c) View 96 and 252
    Deviation analysis results
    Other models and multi-view point cloud registration results. The size of the four objects : (a) 50 mm70 mm; (b) 53 mm75 mm; (c) 63 mm60 mm; (d) 50 mm50 mm其他模型及多视角点云配准结果图。四种物体的尺寸:(a) 50 mm70 mm;(b) 53 mm75 mm:(c) 63 mm60 mm;(d) 50 mm50 mm
    • Table 1. Experimental results of adjacent point cloud registration

      View table
      View in Article

      Table 1. Experimental results of adjacent point cloud registration

      Registration method Average FitnessAverage RmseFailed/ times Time/s
      ICP[4]0.8760.12086.528
      Generalized ICP[5]0.8790.121810.849
      FPFH-based method[15]0.8960.106187.156
      MRA0.8980.10608.868
    • Table 2. Experimental results of adjacent point cloud registration (View 37 & View 38)

      View table
      View in Article

      Table 2. Experimental results of adjacent point cloud registration (View 37 & View 38)

      Registration methodFitnessRmseTime/s
      ICP[4]0.3620.1610.130
      Generalized ICP[5]0.3500.1630.184
      FPFH-based method[15]0.5800.0950.552
      MRA0.5800.0950.795
    • Table 3. Experimental results of adjacent point cloud registration (View 210 & View 211)

      View table
      View in Article

      Table 3. Experimental results of adjacent point cloud registration (View 210 & View 211)

      Registration methodFitnessRmseTime/s
      ICP[4]0.7320.1320.038
      Generalized ICP[5]0.7280.1320.091
      FPFH-based method[15]0.5150.1810.226
      MRA0.7320.1320.020
    • Table 4. Parameter of multi-view point cloud registration experiment

      View table
      View in Article

      Table 4. Parameter of multi-view point cloud registration experiment

      Registration methodDistance constraintsReal-time optimizationLoop closure detection method
      Ref.[15] method ΟΟFPFH-based method
      Proposed methodΠΠMRA
    • Table 5. Results of multi-view point cloud registration experiment

      View table
      View in Article

      Table 5. Results of multi-view point cloud registration experiment

      Registration methodDrift alleviatedTime/s
      Ref.[15] method Ο174.015
      Proposed methodΠ12.375
    • Table 6. Structure of two methods’ pose graph

      View table
      View in Article

      Table 6. Structure of two methods’ pose graph

      Registration methodVerticesOdometry edgesLoop closure edges
      Ref.[15] method 4504491138
      Proposed method450449335
    Tools

    Get Citation

    Copy Citation Text

    Fei Liu, Hanlin Huang, Tian Yang, Wenbo Li, Yang Yang. Robust multi-view registration method for narrow scenes[J]. Infrared and Laser Engineering, 2022, 51(12): 20220114

    Download Citation

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

    Category: Photoelectric measurement

    Received: Feb. 21, 2022

    Accepted: Apr. 1, 2022

    Published Online: Jan. 10, 2023

    The Author Email:

    DOI:10.3788/IRLA20220114

    Topics