Journal of Applied Optics, Volume. 43, Issue 1, 100(2022)

Design and research of bridge cracks detection method based on Mask RCNN

Yanna LIAO and Danyang DOU*
Author Affiliations
  • School of Electronic Engineering, Xi'an University of Posts and Telecommunications, Xi'an 710121, China
  • show less
    Figures & Tables(14)
    Network structure of Mask RCNN
    Network structure of improved FPN
    Structure diagram of SEM module
    Add feature fusion
    Category classification diagram of crack diseases
    Data set annotation visualization
    Loss curves
    Performance comparison diagram of each network model
    Detection results of fine cracks
    Detection results of larger cracks
    Detection results of complex cracks
    • Table 1. Comparison of performance indicators before and after network improvement of strategy 1

      View table
      View in Article

      Table 1. Comparison of performance indicators before and after network improvement of strategy 1

      平均精度 均值 裂缝平均 精度 检测速度/ (image/s)
      Mask RCNN0.8350.8355.000
      Mask RCNN_SEM0.9010.9016.065
    • Table 2. Comparison of performance indicators before and after network improvement of strategy 2

      View table
      View in Article

      Table 2. Comparison of performance indicators before and after network improvement of strategy 2

      平均精度 均值 裂缝平均 精度 破损平均 精度 检测速度 (image/s)
      Mask RCNN_FPN0.8870.9240.8505.000
      Mask RCNN_P_FPN0.9440.9480.9406.065
    • Table 3. Comparison of performance indicators of network model based on ResNet101 and ResNet50 respectively

      View table
      View in Article

      Table 3. Comparison of performance indicators of network model based on ResNet101 and ResNet50 respectively

      检测 准确率 平均精度 均值 检测速度/ (image/s)
      Mask RCNN_ResNet101+SEM0.9980.9446.065
      Mask RCNN_ResNet50+SEM0.8500.7945.122
    Tools

    Get Citation

    Copy Citation Text

    Yanna LIAO, Danyang DOU. Design and research of bridge cracks detection method based on Mask RCNN[J]. Journal of Applied Optics, 2022, 43(1): 100

    Download Citation

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

    Category: OPTICAL METROLOGY AND MEASUREMENT

    Received: Aug. 12, 2021

    Accepted: --

    Published Online: Mar. 7, 2022

    The Author Email: Danyang DOU (212335154@qq.com)

    DOI:10.5768/JAO202243.0103005

    Topics