Laser & Optoelectronics Progress, Volume. 61, Issue 22, 2237003(2024)

Road Information Extraction Method Based on Hypervoxel Segmentation

Zhe Su1, Li Yang2, Zai Luo1、*, Wensong Jiang1, and Hongmei Fang3
Author Affiliations
  • 1College of Metrology Measurement and Instrument, China Jiliang University, Hangzhou 310018, Zhejiang , China
  • 2College of Information Engineering, China Jiliang University, Hangzhou 310018, Zhejiang , China
  • 3Hangzhou Mutual Recognition Quality Technology Research Institute, Hangzhou 310015, Zhejiang , China
  • show less
    Figures & Tables(11)
    Overall flowchart
    Schematic diagram of boundary point subdivision
    Schematic diagram of boundary point extraction
    Self-collecting data experiment scene. (a) Experimental scene; (b) experimental platform details
    Rough extraction of ground points in IQTM point cloud dataset. (a) Primary point cloud; (b) point cloud after filtering non-ground point
    Registration results of road edge. (a) Target point cloud; (b) registration result by VCCS; (c) registration result by VCCS-kNN; (d) registration result by proposed method
    Extraction results of IQTM dataset. (a) Extraction result of boundary point; (b) extraction result of road edge
    Self-collected dataset. (a) Self-collected data; (b) hypervoxel segmentation result; (c) extraction result of boundary point; (d) extraction result of road edge
    Extraction results of lane line. (a) IQTM data driving area; (b) extraction result of IQTM data by global threshold method; (c) extraction result of IQTM data by proposed method; (d) self-collected data driving area; (e) extraction result of self-collected data by global threshold method; (f) extraction result of self-collected data by proposed method
    • Table 1. Evaluation results for different methods

      View table

      Table 1. Evaluation results for different methods

      MethodPoint cloud loss rateBoundary recall rateUndersegmentation error
      VCCS0.6639.217.9
      VCCS-kNN0.3543.714.8
      Proposed method058.38.3
    • Table 2. Evaluation of lane line extraction results by proposed method

      View table

      Table 2. Evaluation of lane line extraction results by proposed method

      Sample areaFit area /m2Actual area /m2Deviation degree /%Fit distance /mActual distance /mOffset distance /m
      10.5740.562.53.303.270.03
      20.5710.562.03.353.310.04
      30.5870.62.23.273.250.02
      40.5640.582.73.293.330.04
      50.4140.433.73.333.370.04
    Tools

    Get Citation

    Copy Citation Text

    Zhe Su, Li Yang, Zai Luo, Wensong Jiang, Hongmei Fang. Road Information Extraction Method Based on Hypervoxel Segmentation[J]. Laser & Optoelectronics Progress, 2024, 61(22): 2237003

    Download Citation

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

    Category: Digital Image Processing

    Received: Dec. 22, 2023

    Accepted: Mar. 25, 2024

    Published Online: Nov. 15, 2024

    The Author Email: Zai Luo (luozai@cjlu.edu.cn)

    DOI:10.3788/LOP232716

    CSTR:32186.14.LOP232716

    Topics