Acta Optica Sinica, Volume. 44, Issue 9, 0911001(2024)

ECT Image Reconstruction Based on Fuzzy Mode Recognition and Sensitive Field Optimization

Guoxing Huang1, Chao Li1, Zhenhua Wu1, Jingwen Wang1、*, Taoya Yuan2, and Weidang Lu1
Author Affiliations
  • 1School of Information Engineering, Zhejiang University of Technology, Hangzhou 310023, Zhejiang, China
  • 2School of Information Science and Engineering, Harbin Institute of Technology, Weihai 264209, Shandong, China
  • show less
    Figures & Tables(14)
    ECT system structure diagram
    Distributions of sensitivity field under different flow patterns. (a) Full pipe; (b) empty pipe; (c) annular; (d) core
    Schematic diagram of pixel division of imaging area
    Flow chart for flow pattern identification
    Schematic diagram of the extension method for sensitive domains under feature extraction
    COMSOL simulation model of ECT system
    Initial reconstructed images and original flow patterns of empty and full tubes (corresponding to the flow pattern identification results in Table 2)
    Reconstructed images of different algorithms in simulation
    Image relative error histogram of different algorithms
    Histogram of image correlation coefficients for different algorithms
    • Table 1. Statistical results of ECT flow pattern identification based on fuzzy mode

      View table

      Table 1. Statistical results of ECT flow pattern identification based on fuzzy mode

      Flow patternIdentification accuracy /%
      Noise free experimentNoise is 60 dBNoise is 40 dB
      1/3 laminar100100100
      1/2 laminar100100100
      2/3 laminar10010098
      Annular100100100
      Core100100100
      Bubble1009892
      Empty pipe100100100
      Full pipe100100100
    • Table 2. Fuzzy characteristic parameters of ECT flow pattern identification based on fuzzy mode

      View table

      Table 2. Fuzzy characteristic parameters of ECT flow pattern identification based on fuzzy mode

      Experimental flow patternFuzzy feature parameter valueValue of affiliationFlow pattern identification result
      x1x2x3ϖY1ϖY2ϖY3
      1/3 laminar0.1730.3430.0570.4510.5670.4061/3 laminar[Fig. 7(a)]
      1/2 laminar0.3350.5840.1140.3500.4470.2731/2 laminar[Fig. 7(b)]
      Annular0.3040.0860.2340.4710.1940.527Annular[Fig. 7(c)]
      Core0.2670.0010.4550.5320.1650.753Core[Fig. 7(d)]
      2/3 laminar0.6310.3560.3650.2800.2610.2662/3 laminar[Fig. 7(e)]
      Bubble0.5010.0940.5300.4960.1880.494Bubble[Fig. 7(f)]
      Full pipe0.7880.002Full pipe[Fig. 7(g)]
      Empty pipe0.0030.0050.004Empty pipe[Fig. 7(h)]
    • Table 3. Relative error of image reconstruction by five algorithms

      View table

      Table 3. Relative error of image reconstruction by five algorithms

      AlgorithmRelative error
      1/3 laminar1/2 laminar2/3 laminarAnnularCoreBubble
      Landweber0.2460.2600.1920.2820.4500.443
      Tikhonov0.5190.5490.3720.3930.4050.438
      Kalman0.3580.3990.2730.2560.7590.757
      CGLS0.3890.4000.2860.3660.6810.925
      Proposed method0.1780.1150.0840.1830.3000.406
    • Table 4. Correlation coefficients of reconstructed images by five algorithms

      View table

      Table 4. Correlation coefficients of reconstructed images by five algorithms

      AlgorithmCorrelation coefficient
      1/3 laminar1/2 laminar2/3 laminarAnnularCoreBubble
      Landweber0.9150.8860.8330.7900.7990.679
      Tikhonov0.6950.4340.3070.6500.8320.680
      Kalman0.8350.7820.7400.8410.6660.190
      CGLS0.8400.8240.7490.7130.7610.189
      Proposed method0.9410.9490.9280.9020.8950.699
    Tools

    Get Citation

    Copy Citation Text

    Guoxing Huang, Chao Li, Zhenhua Wu, Jingwen Wang, Taoya Yuan, Weidang Lu. ECT Image Reconstruction Based on Fuzzy Mode Recognition and Sensitive Field Optimization[J]. Acta Optica Sinica, 2024, 44(9): 0911001

    Download Citation

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

    Category: Imaging Systems

    Received: Jan. 8, 2024

    Accepted: Feb. 19, 2024

    Published Online: May. 15, 2024

    The Author Email: Wang Jingwen (wangjingwenhappy@126.com)

    DOI:10.3788/AOS240452

    CSTR:32393.14.AOS240452

    Topics