Journals Articles Conferences News About CLP
Submit a paper Email Alert
Search
Search
Articles
Journals
News
Advanced Search
Top Searches
2D MaterialsTransformation opticsQuantum PhotonicsSmall lasersSemiconductor UV PhotonicsSupersymmetric microring laser arrays

Laser & Optoelectronics Progress, Volume. 59, Issue 12, 1215010(2022)

Fringe Segmentation Algorithm Based on Improved U-Net

Wenwei Yan1,2,3,4, Shuai Chen1,2,4、*, Baoyan Mu1,2,4, and Liang Gao1,2,4
Author Affiliations
  • 1Shenyang Institute of Automation, Chinese Academy of Sciences, Shenyang 110016, Liaoning , China
  • 2Institutes for Robotics and Intelligent Manufacturing, Chinese Academy of Sciences, Shenyang 110169, Liaoning , China
  • 3University of Chinese Academy of Sciences, Beijing 100049, China
  • 4Key Laboratory on Intelligent Detection and Equipment Technology of Liaoning Province, Shenyang 110179, Liaoning , China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (18)
    Equations (16)
    References (20)
    Cited By (2)
    Tools
    Paper Information
    Topics
    Figures & Tables(18)
    Image acquisition system

    Fig. 1. Image acquisition system

    Download full size
    Captured image

    Fig. 2. Captured image

    Download full size
    Marked map

    Fig. 3. Marked map

    Download full size
    Flow chart of image segmentation

    Fig. 4. Flow chart of image segmentation

    Download full size
    U-Net network

    Fig. 5. U-Net network

    Download full size
    Coordinate attention mechanism

    Fig. 6. Coordinate attention mechanism

    Download full size
    Pyramid pooling module

    Fig. 7. Pyramid pooling module

    Download full size
    VGG16 network

    Fig. 8. VGG16 network

    Download full size
    Improved U-Net network

    Fig. 9. Improved U-Net network

    Download full size
    Training loss value and accuracy value

    Fig. 10. Training loss value and accuracy value

    Download full size
    Validation loss value and accuracy value

    Fig. 11. Validation loss value and accuracy value

    Download full size
    Segmentation results of different algorithms

    Fig. 12. Segmentation results of different algorithms

    Download full size
    Fringe segmentation of improved U-Net algorithm in complex scene. (a) Original image; (b) ground true; (c) segmentation result of improved U-Net algorithm

    Fig. 13. Fringe segmentation of improved U-Net algorithm in complex scene. (a) Original image; (b) ground true; (c) segmentation result of improved U-Net algorithm

    Download full size
    Extraction of light stripe feature points corresponding to stiffeners in different directions. (a) Feature point extraction of horizontal stiffener stripes; (b) feature point extraction of inclined stiffener stripes; (c) feature point extraction of vertical stiffener stripes

    Fig. 14. Extraction of light stripe feature points corresponding to stiffeners in different directions. (a) Feature point extraction of horizontal stiffener stripes; (b) feature point extraction of inclined stiffener stripes; (c) feature point extraction of vertical stiffener stripes

    Download full size
    Metal workpiece

    Fig. 15. Metal workpiece

    Download full size

    • Table 1. Index values of different algorithms

      View table

      Table 1. Index values of different algorithms

      AlgorithmmIoU /%mpa /%
      Tradition algorithm49.4372.68
      U-Net84.2893.59
      VGG16+U-Net84.3993.8
      Attention U-Net84.2994.45
      VGG16+ Attention U-Net86.5594.88
      PSPNet88.3294.95
      ENet87.3294.09
      Proposed algorithm89.7395.61

    • Table 2. Measurement results of stiffeners in different directions of metal workpiece

      View table

      Table 2. Measurement results of stiffeners in different directions of metal workpiece

      Measuring positionL /mml /mmΔ /mmε /%
      a10.029.95338-0.0666-0.665
      b9.989.91998-0.06002-0.601
      c9.969.91839-0.04161-0.418
      d10.009.92784-0.07216-0.722

    • Table 3. Repeatability measurement of stiffener at position a of metal workpiece

      View table

      Table 3. Repeatability measurement of stiffener at position a of metal workpiece

      Measuring positionnl /mmlave /mmσδ /%
      a19.953389.947650.013470.14
      29.93935
      39.96358
      49.93430
    Tools

    Get Citation

    Copy Citation Text

    Wenwei Yan, Shuai Chen, Baoyan Mu, Liang Gao. Fringe Segmentation Algorithm Based on Improved U-Net[J]. Laser & Optoelectronics Progress, 2022, 59(12): 1215010

    Download Citation

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

    Category: Machine Vision

    Received: Jul. 8, 2021

    Accepted: Aug. 17, 2021

    Published Online: May. 23, 2022

    The Author Email: Shuai Chen (chenshuai@sia.cn)

    DOI:10.3788/LOP202259.1215010

    Topics

    laser devices and laser physics
    Lasers and Laser Optics
    Laser physics
    laser manufacturing
    Instrumentation, Measurement and Metrology