Acta Optica Sinica, Volume. 43, Issue 14, 1412001(2023)

Large Deformation Measurement Method of Speckle Images Based on Deep Learning

Hong Xiao, Chengnan Li, and Mingchi Feng*
Author Affiliations
  • School of Advanced Manufacturing Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China
  • show less
    Figures & Tables(12)
    Four kinds of typical speckle images. (a) Speckle image Ⅰ; (b) speckle image Ⅱ; (c) speckle image Ⅲ; (d) speckle image Ⅳ
    Partial samples of speckle displacement field image dataset
    Convolutional block attention module. (a) Channel attention module; (b) sapatial attention module; (c) convolutional block attention module
    Two different convolution modules. (a) General convolution; (b) depthwise separable convolution
    Structure diagram of DICNet
    Loss decline curves of FlowNetS, FlowNetC, and UNet during training. (a) Loss of training set; (b) loss of validation set
    Performance of algorithms and models on test set. (a) Sample 1; (b) sample 2; (c) sample 3; (d) sample 4
    Performance of algorithms and models on DIC challenge dataset. (a) Sample 8; (b) sample 10; (c) sample 12; (d) sample 17
    • Table 1. Generation parameters of classⅡ and Ⅳ speckle images

      View table

      Table 1. Generation parameters of classⅡ and Ⅳ speckle images

      ClassRspeckle /pixelNspeckleΓTspeckleNimage
      0.51500000.6E200
      0.71500000.5E200
      1100000.5P200
      360000.6P200
      0.71500000.5U100
      1.5500000.6L100
      1500000.7E100
      2200000.8P100
    • Table 2. Model information and training results of FlowNetS, FlowNetC, and UNet

      View table

      Table 2. Model information and training results of FlowNetS, FlowNetC, and UNet

      ModelNparams /MTtrain /minLMSEtr /pixel2LMSEval /pixel2Tinference /ms
      FlowNetC39.1751470.3580.4778.369
      FlowNetS38.675800.4260.2935.272
      UNet7.760990.0560.0735.532
    • Table 3. Performance of DICNet on self-built training set and validation set

      View table

      Table 3. Performance of DICNet on self-built training set and validation set

      Algorithm

      and model

      RMSE on training set LRMSEtr /pixelRMSE on validation set LRMSEval /pixel

      Mean time

      Tinference /ms

      MeanMaximumMinimumStandard deviationMeanMaximumMinimum

      Standard

      deviation

      Ncorr0.5121.3090.0560.2410.5191.2450.0460.2398089.898
      OpenCorr-CPU0.3072.8970.0040.1930.3171.6190.0360.19616492.512
      OpenCorr-GPU0.2702.8770.0030.1710.2751.2760.0130.1632383.299
      DeepDIC0.1740.3250.0970.0190.1740.3630.0890.01915.800
      StrainNet0.1870.7840.0550.0710.1830.7460.0540.0663.727
      UNet0.0920.2590.0490.0260.0910.2410.0470.0264.015
      DICNet0.0560.2090.0320.0190.0550.1890.0340.0184.919
    • Table 4. Performance of DICNet on self-built test set

      View table

      Table 4. Performance of DICNet on self-built test set

      Algorithm

      and model

      RMSE on test set LRMSEte /pixel

      Number of model parameter

      Nparams /M

      MeanMaximumMinimumStandard deviation
      Ncorr0.5091.1270.0710.223
      OpenCorr-CPU0.1650.5960.0290.091
      OpenCorr-GPU0.1760.7320.0260.109
      DeepDIC0.1830.2500.1400.01949.970
      StrainNet0.2060.5570.0670.07238.663
      UNet0.1740.4130.0700.0507.760
      DICNet0.1580.2670.0780.0396.398
    Tools

    Get Citation

    Copy Citation Text

    Hong Xiao, Chengnan Li, Mingchi Feng. Large Deformation Measurement Method of Speckle Images Based on Deep Learning[J]. Acta Optica Sinica, 2023, 43(14): 1412001

    Download Citation

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

    Category: Instrumentation, Measurement and Metrology

    Received: Dec. 2, 2022

    Accepted: Mar. 20, 2023

    Published Online: Jul. 13, 2023

    The Author Email: Mingchi Feng (fengmc@cqupt.edu.cn)

    DOI:10.3788/AOS222084

    Topics