• Optics and Precision Engineering
  • Vol. 31, Issue 3, 340 (2023)
Kaiyuan GAO1, Lei LIU1, Haihua CUI1,*, Pengcheng LI1..., Xiaoxu LIU2 and Lin LIU2|Show fewer author(s)
Author Affiliations
  • 1College of Mechanical & Electrical Engineering, Nanjing University of Aeronautics and Astronautics, Nanjing2006, China
  • 2Beijing Institute of Aerospace Metrology and Testing Technology, Beijing100076, China
  • show less
    DOI: 10.37188/OPE.20233103.0340 Cite this Article
    Kaiyuan GAO, Lei LIU, Haihua CUI, Pengcheng LI, Xiaoxu LIU, Lin LIU. Multi-scale decomposition of point cloud data based on wavelet transform[J]. Optics and Precision Engineering, 2023, 31(3): 340 Copy Citation Text show less
    Schematic diagram of multi-scale 3D measurements
    Fig. 1. Schematic diagram of multi-scale 3D measurements
    Multilevel wavelet decomposition space
    Fig. 2. Multilevel wavelet decomposition space
    Convolutional nuclei
    Fig. 3. Convolutional nuclei
    Grid convolution
    Fig. 4. Grid convolution
    Macro-micro combined measurement results of blade
    Fig. 5. Macro-micro combined measurement results of blade
    Decomposition results of wavelet method
    Fig. 6. Decomposition results of wavelet method
    Decomposation results of random sampling method
    Fig. 7. Decomposation results of random sampling method
    Decomposation results of curvature sampling method
    Fig. 8. Decomposation results of curvature sampling method
    Results of scale spatial decomposition method
    Fig. 9. Results of scale spatial decomposition method
    Comparison of differences between different decomposition methods
    Fig. 10. Comparison of differences between different decomposition methods
    Pore registration results
    Fig. 11. Pore registration results
    Macro-micro combined measurement results of raster
    Fig. 12. Macro-micro combined measurement results of raster
    Grid edge registration results
    Fig. 13. Grid edge registration results
    Inspection results of leaf without pores
    Fig. 14. Inspection results of leaf without pores
    Leaf edge registration results
    Fig. 15. Leaf edge registration results
    点云点个数均方差/mm面维数面维数差体维数体维数差
    原始1 073 59101.156 10-0.364 50
    栅格1 073 59101.144 80.011 3-0.362 80.001 7
    采样1次129 5280.002 570 21.108 10.048 0-0.369 10.004 6
    采样2次15 3930.005 718 01.081 40.074 7-0.391 10.026 6
    采样3次1 8720.012 060 81.017 40.138 7-0.405 20.040 7
    采样4次2160.023 510 10.945 90.210 2-0.640 20.275 7
    Table 1. Statistic of wavelet method decomposition results
    点云点个数均方差/mm面维数面维数差体维数体维数差
    原始1 073 59101.156 10-0.364 50
    分解1次128 83101.146 80.009 3-0.372 80.008 3
    分解2次15 46001.072 60.083 5-0.392 10.027 6
    分解3次1 85501.012 10.144 0-0.425 90.061 4
    Table 2. Statistics of random sampling decomposition results
    点云点个数均方差/mm面维数面维数差体维数体维数差
    原始1 073 59101.156 10-0.364 50
    分解1次128 83101.068 00.088 1-0.353 90.010 6
    分解2次15 46001.030 70.125 4-0.366 30.011 8
    分解3次1 84200.995 50.160 6-0.442 20.077 7
    Table 3. Statistics of curvature sampling decomposition results
    点云点个数均方差/mm面维数面维数差体维数体维数差
    原始1 073 59101.156 10-0.364 50
    分解1次128 8310.002 075 841.091 10.065 0-0.407 80.043 3
    分解2次15 4600.002 070 131.081 10.075 0-0.431 60.067 1
    分解3次1 8550.002 014 681.008 80.147 3-0.512 30.147 8
    Table 4. Statistics of scale spatial decomposition results
    分解次数配准时间/s配准均方根误差/mm
    0(原始数据)301.9860.042 275 9
    1114.6380.023 725 1
    239.1650.017 195 0
    329.3680.016 334 5
    Table 5. Comparison of multiple decomposition registration error for wavelet method
    分解次数配准时间/s配准均方根误差/mm
    0(原始数据)207.1640.051 801
    147.9430.030 997
    220.0530.029 199
    313.6220.029 079
    Table 6. Approximate scale of raster constrains the comparison of registration errors
    分解次数配准时间/s配准均方根误差/mm
    0(原始数据)215.6830.016 210 9
    1147.6740.016 210 7
    230.5860.010 027 6
    315.6050.008 333 7
    Table 7. Comparison of approximate scale constraint registration errors of blade edges
    Kaiyuan GAO, Lei LIU, Haihua CUI, Pengcheng LI, Xiaoxu LIU, Lin LIU. Multi-scale decomposition of point cloud data based on wavelet transform[J]. Optics and Precision Engineering, 2023, 31(3): 340
    Download Citation