Laser & Optoelectronics Progress, Volume. 58, Issue 6, 610015(2021)

Point Cloud Filtering Algorithm Based on Image Processing

Zhang Jianmin1, Chen Fujian2, and Long Jiale1、*
Author Affiliations
  • 1Faculty of Intelligent Manufacturing, Wuyi University, Jiangmen, Guangdong 529020, China
  • 2School of Electronic and Information Engineering, Shenzhen University, Shenzhen, Guangdong 518060, China
  • show less
    Figures & Tables(19)
    Object 1 to be measured
    Front view of point cloud
    Side view of point cloud
    Schematic of point cloud filtering algorithm based on image processing
    Flow chart of point cloud filtering algorithm based on image processing
    Correspondence diagram of point cloud coordinates, image coordinates, and point cloud number
    Schematic of image processing (selecting the largest connected area)
    Schematic of image processing (selecting the connected area with the largest number of point clouds)
    Filtering process of point cloud for object 1. (a) Result of k nearest neighbor filtering; (b) point cloud mapped image; (c) image closing operation; (d) image filtering processing; (e) result of point cloud filtering (side view); (f) result of point cloud filtering (front view)
    Filtering process of point cloud for object 2. (a) Object 2 to be measured; (b) front view of point cloud; (c) side view of point cloud; (d) result of k nearest neighbor filtering; (e) point cloud mapped image; (f) image closing operation; (g) image filtering processing; (h) result of point cloud filtering (front view); (i) result of point cloud filtering (side view)
    Filtering process of point cloud for object 3. (a) Object 3 to be measured; (b) front view of point cloud; (c) side view of point cloud; (d) result of k nearest neighbor filtering; (e) point cloud mapped image; (f) image closing operation; (g) image filtering processing; (h) result of point cloud filtering (front view); (i) result of point cloud filtering (side view)
    Filtering results of different algorithms for object 1. (a) Proposed algorithm; (b) k nearest neighbor filtering algorithm; (c) radius filtering algorithm; (d) voxel filtering algorithm
    Filtering results of different algorithms for object 2. (a) Proposed algorithm; (b) k nearest neighbor filtering algorithm; (c) radius filtering algorithm; (d) voxel filtering algorithm
    Filtering results of different algorithms for object 3. (a) Proposed algorithm; (b) k nearest neighbor filtering algorithm; (c) radius filtering algorithm; (d) voxel filtering algorithm
    • Table 1. Time consuming of largest connected area method

      View table

      Table 1. Time consuming of largest connected area method

      Model nameNumber of raw point cloudsNumber of filtered point cloudsTime consuming in knearest neighbor filtering /sTime consuming in remaining processing /sTotal time /s
      Object 11889251677960.53960.36890.9085
      Object 253838531720.11810.06110.1792
      Object 364933315280.16160.26650.4281
    • Table 2. Time consuming of the connected area with the largest number of point clouds

      View table

      Table 2. Time consuming of the connected area with the largest number of point clouds

      Model nameNumber of raw point cloudsNumber of filtered point cloudsTime consuming in k nearest neighbor filtering /sTime consuming in remaining processing /sTotal time /s
      Object 11889251677960.53960.41200.9516
      Object 253838531720.11810.06150.1796
      Object 364933315280.16160.27430.4359
    • Table 3. Filtering effect analysis of object 1

      View table

      Table 3. Filtering effect analysis of object 1

      ParameterProposed algorithmk nearest neighbor filtering algorithmRadius filtering algorithmVoxel filteringalgorithm
      Number of filtered point clouds16779518702818794090878
      Time consuming in filtering /s0.9090.4860.7550.081
      NTP16773616779616785712702
      NFP60192322008378176
      P /%99.96489.71789.31413.977
    • Table 4. Filtering effect analysis of object 2

      View table

      Table 4. Filtering effect analysis of object 2

      ParameterProposed algorithmk nearest neighbor filtering algorithmRadius filtering algorithmVoxel filtering algorithm
      Number of filtered point clouds53172534175363724428
      Time consuming in filtering /s0.3410.1340.2130.025
      NTP5316653172531783518
      NFP624545920910
      P /%99.98899.54199.14414.402
    • Table 5. Filtering effect analysis of object 3

      View table

      Table 5. Filtering effect analysis of object 3

      ParameterProposed algorithmk nearest neighbor filtering algorithmRadius filtering algorithmVoxel filtering algorithm
      Number of filtered point clouds31528512946391561061
      Time consuming in filtering /s0.4270.1650.2730.028
      NTP31429315283162927262
      NFP99197663228633799
      P /%99.68661.46549.48644.647
    Tools

    Get Citation

    Copy Citation Text

    Zhang Jianmin, Chen Fujian, Long Jiale. Point Cloud Filtering Algorithm Based on Image Processing[J]. Laser & Optoelectronics Progress, 2021, 58(6): 610015

    Download Citation

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

    Category: Image Processing

    Received: Aug. 12, 2020

    Accepted: --

    Published Online: Mar. 11, 2021

    The Author Email: Jiale Long (longjiale_528@126.com)

    DOI:10.3788/LOP202158.0610015

    Topics