• Acta Photonica Sinica
  • Vol. 54, Issue 1, 0106002 (2025)
Miao YU1, Yutong HE2, Tianying CHANG1, Hongliang CUI1,*..., Suihhu DANG3, Liangping XIA3, Liming LIU2, Zichuan YI2, Xinjian PAN2 and Qingguo GAO2|Show fewer author(s)
Author Affiliations
  • 1Shenzhen Institute of Advanced Technology,Chinese Academy of Sciences,Shenzhen 518000,China
  • 2School of Electronic Information Engineering,University of Electronic Science and Technology of China,Zhongshan Institute,Zhongshan 528402,China
  • 3School of Electronic and Information Engineering,Yangtze Normal University,Chongqing 408000,China
  • show less
    DOI: 10.3788/gzxb20255401.0106002 Cite this Article
    Miao YU, Yutong HE, Tianying CHANG, Hongliang CUI, Suihhu DANG, Liangping XIA, Liming LIU, Zichuan YI, Xinjian PAN, Qingguo GAO. Distributed Fiber Optic Acoustic Sensing System Based on Fading Mask Autoencoder and Application in Water Navigation Security Events Identification[J]. Acta Photonica Sinica, 2025, 54(1): 0106002 Copy Citation Text show less
    DAS system structure and system physical diagram
    Fig. 1. DAS system structure and system physical diagram
    The DAS system acquisition contains the fading signal
    Fig. 2. The DAS system acquisition contains the fading signal
    Single-time differential result of amplitude information
    Fig. 3. Single-time differential result of amplitude information
    Grubbs criteria to judge the fading position information
    Fig. 4. Grubbs criteria to judge the fading position information
    Fading mask strategy flow of FMAE
    Fig. 5. Fading mask strategy flow of FMAE
    FMAE model structure
    Fig. 6. FMAE model structure
    Simulation experiment of water navigation security event
    Fig. 7. Simulation experiment of water navigation security event
    Data visualization of the four types of ship events in waters
    Fig. 8. Data visualization of the four types of ship events in waters
    Mean comparison of ten cross-validation training results for two models:FMAE and MAE
    Fig. 9. Mean comparison of ten cross-validation training results for two models:FMAE and MAE
    Comparison of training results for the four models,FMAE/ViT/SENet/ResNet
    Fig. 10. Comparison of training results for the four models,FMAE/ViT/SENet/ResNet
    Confusion matrix of the four models
    Fig. 11. Confusion matrix of the four models
    Event type(I)drone ship(II)boating(III)speedboat(IV)no eventTotal
    Train set16 60018 00017 00018 00069 600
    Test set2 0002 0002 0002 0008 000
    Total18 60020 00019 00020 00077 600
    Table 1. Data sets used for the four types of shipping events in waters

    Average training

    accuracy/%

    Average training

    loss

    Average test

    accuracy/%

    Average precisionAverage F1-score
    FMAE98.340.109 493.010.934 50.929 6
    MAE93.440.204 486.560.893 10.858 7
    Table 2. Comparison of performance measures between FMAE and MAE models
    IIIIIIIV
    RecallAccuracy
    FMAE0.786 00.944 50.990 01.000 00.930 1
    ViT0.379 50.994 51.000 01.000 00.843 5
    SENet0.576 50.946 51.000 01.000 00.880 7
    ResNet0.707 00.772 51.000 01.000 00.869 8
    PrecisionAverage precision
    FMAE0.934 10.818 80.994 90.990 10.934 5
    ViT0.985 70.626 00.974 11.000 00.896 4
    SENet0.915 00.711 90.961 01.000 00.897 0
    ResNet0.769 30.727 40.997 00.984 20.869 4
    F1-scoreAverage F1-score
    FMAE0.853 60.877 20.992 50.995 00.929 6
    ViT0.548 00.768 30.986 91.000 00.825 8
    SENet0.707 30.812 60.980 11.000 00.875 0
    ResNet0.736 80.749 20.998 50.992 00.869 1
    Table 3. Performance indicators of the four models,FMAE / ViT / SENet / ResNet
    FMAEViTSENetResNet
    Inference time10.75 ms6.98 ms16.96 ms17.95 ms
    Table 4. Inference time of four models,FMAE / ViT / SENet / ResNet
    Miao YU, Yutong HE, Tianying CHANG, Hongliang CUI, Suihhu DANG, Liangping XIA, Liming LIU, Zichuan YI, Xinjian PAN, Qingguo GAO. Distributed Fiber Optic Acoustic Sensing System Based on Fading Mask Autoencoder and Application in Water Navigation Security Events Identification[J]. Acta Photonica Sinica, 2025, 54(1): 0106002
    Download Citation