Laser & Optoelectronics Progress, Volume. 57, Issue 16, 161017(2020)

Recognition Method of Waste Non-Ferrous Metal Fragments Based on Machine Vision

Zhenyuan Zhang, Xunpeng Qin*, and Yifeng Li
Author Affiliations
  • Hubei Key Laboratory of Advanced Technology for Automotive Component, School of Automotive Engineering, Wuhan University of Technology, Wuhan, Hubei 430070, China
  • show less
    Figures & Tables(11)
    Experimental platform of sorting metal fragments
    Flow chart of extracting contour of metal fragments[13]
    Extracting contours of metal fragments. (a) Without information filtering; (b) with information filtering
    Extracting window pictures of metal fragments
    Scatterplot of metal fragment texture feature data
    Histograms of metal fragment color feature. (a) Aluminum; (b) copper; (c) steel
    Curve of cumulative explained variance contribution rate
    Classification results of cross-validation
    • Table 1. Texture feature data of metal fragments

      View table

      Table 1. Texture feature data of metal fragments

      MaterialCoarsenessContrastDirectionality
      Copperfragment18.517.515.916.215.510.45.74.713.38.730.028.667.837.929.5
      Aluminumfragment15.114.816.318.215.728.435.238.624.947.919.820.419.932.219.3
      Steelfragment15.616.517.415.215.010.912.913.46.59.934.437.448.165.645.9
    • Table 2. Comparison of kernel function accuracy rate

      View table

      Table 2. Comparison of kernel function accuracy rate

      Kernel functionAccuracy rate /%
      Linear78.89
      Poly80.89
      Radial basis function89.11
      Sigmoid87.33
    • Table 3. Performance comparison of different algorithms

      View table

      Table 3. Performance comparison of different algorithms

      AlgorithmAccuracy rate /%Detection time /s
      PCA+SVM93.890.51
      SVM94.110.98
      Random Forest91.111.21
      KNN90.561.25
    Tools

    Get Citation

    Copy Citation Text

    Zhenyuan Zhang, Xunpeng Qin, Yifeng Li. Recognition Method of Waste Non-Ferrous Metal Fragments Based on Machine Vision[J]. Laser & Optoelectronics Progress, 2020, 57(16): 161017

    Download Citation

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

    Category: Image Processing

    Received: Nov. 14, 2019

    Accepted: Jan. 16, 2020

    Published Online: Aug. 5, 2020

    The Author Email: Xunpeng Qin (qxp915@hotmail.com)

    DOI:10.3788/LOP57.161017

    Topics