Journal of Applied Optics, Volume. 44, Issue 1, 86(2023)

Image segmentation method of surface defects for metal workpieces based on improved U-net

Yi WANG1...2, Xiaojie GONG1,*, and Hao SU13 |Show fewer author(s)
Author Affiliations
  • 1College of Electrical Engineering, North China University of Science and Technology, Tangshan 063210, China
  • 2Tangshan Technology Innovation Center of Intellectualization of Metal Component Production Line, Tangshan 063210, China
  • 3Tangshan Key Laboratory of Semiconductor Integrated Circuits, Tangshan 063210, China
  • show less
    Figures & Tables(10)
    Structure diagram of U-net network
    Structure diagram of CBAM network
    Structure diagram of improved U-net network
    Visual platform for image acquisition
    Comparison before and after image preprocessing
    Variation curves of loss values
    Variation curves of mean intersection over union
    Variation curves of accuracy rate
    Comparison of segmentation results
    • Table 1. Comparison of Dice indexes

      View table
      View in Article

      Table 1. Comparison of Dice indexes

      No.ApproachDice
      1PW-U-net0.8463
      2U-net0.8489
      3DO-U-net0.8516
      4Att-U-net0.8512
      5Att-DO-U-net0.8674
    Tools

    Get Citation

    Copy Citation Text

    Yi WANG, Xiaojie GONG, Hao SU. Image segmentation method of surface defects for metal workpieces based on improved U-net[J]. Journal of Applied Optics, 2023, 44(1): 86

    Download Citation

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

    Category: Research Articles

    Received: Mar. 24, 2022

    Accepted: --

    Published Online: Feb. 22, 2023

    The Author Email: GONG Xiaojie (1692994031@qq.com)

    DOI:10.5768/JAO202344.0102004

    Topics