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

Acta Optica Sinica, Volume. 43, Issue 2, 0228001(2023)

Optical Fiber Vibration Sensing Detection with High Accuracy Based on YOLOv5s Model

Kang Xue1,2,3, Kun Liu1,2,3、*, Junfeng Jiang1,2,3, Shuang Wang1,2,3, Tianhua Xu1,2,3, Zhenshi Sun1,2,3, Sichen Li1,2,3, Yuelang Huang1,2,3, Xibo Jin1,2,3, and Tiegen Liu1,2,3
Author Affiliations
  • 1School of Precision Instrument and Opto-Electronics Engineering, Tianjin University, Tianjin 300072, China
  • 2Key Laboratory of the Ministry of Education on Optoelectronic Information Technology, Tianjin University, Tianjin 300072, China
  • 3Institute of Optical Fiber Sensing, Tianjin University, Tianjin 300072, China
    • show less
    AbstractGet PDF(in Chinese)
    Figures&Tables (12)
    Equations (0)
    References (21)
    Cited By (5)
    Tools
    Paper Information
    Topics
    Figures & Tables(12)
    Schematic of DMZI-UAV fusion security system. (a) DMZI structure diagram; (b) UAV structure diagram

    Fig. 1. Schematic of DMZI-UAV fusion security system. (a) DMZI structure diagram; (b) UAV structure diagram

    Download full size
    Transition of signal from a time series to a time spectrogram

    Fig. 2. Transition of signal from a time series to a time spectrogram

    Download full size
    Schematic of YOLOv5s network structure

    Fig. 3. Schematic of YOLOv5s network structure

    Download full size
    Neck network diagram

    Fig. 4. Neck network diagram

    Download full size
    Actual experimental drawings. (a) Five typical sensing events: no intrusion, waggling, knocking, crashing, and kicking; (b) outdoor experimental scene diagram: fence and optical fiber cable; (c) QGroundcontrol interface

    Fig. 5. Actual experimental drawings. (a) Five typical sensing events: no intrusion, waggling, knocking, crashing, and kicking; (b) outdoor experimental scene diagram: fence and optical fiber cable; (c) QGroundcontrol interface

    Download full size
    Five typical sensing signals and their corresponding STFT time-frequency diagrams

    Fig. 6. Five typical sensing signals and their corresponding STFT time-frequency diagrams

    Download full size
    STFT diagrams and corresponding UAV images of four typical intrusion events

    Fig. 7. STFT diagrams and corresponding UAV images of four typical intrusion events

    Download full size
    Training results for five typical events. (a) Classification loss; (b) object loss; (c) mAP@0.5; (d) mAP@0.5:0.95

    Fig. 8. Training results for five typical events. (a) Classification loss; (b) object loss; (c) mAP@0.5; (d) mAP@0.5:0.95

    Download full size

    • Table 1. Number of samples selected for the experiment

      View table

      Table 1. Number of samples selected for the experiment

      TypeNo intrusionWagglingKnockingCrashingKicking
      STFT600600600600600
      UAV0700700700700
      Total6001300130013001300

    • Table 2. Experimental parameter table

      View table

      Table 2. Experimental parameter table

      ParameterValue
      Learning rate0.0032
      Momentum0.843
      Weight decay rate（decay）0.00036
      Batch size32

    • Table 3. Comparison of recognition and classification results of three different datasets

      View table

      Table 3. Comparison of recognition and classification results of three different datasets

      ConditionEvent typeP /%R /%F1 /%

      One-dimensional raw data mAP@0.5：0.95

      80.5%

      		Label 182.483.683.0
      Label 281.780.280.9
      Label 382.582.282.3
      Label 484.585.084.7
      Label 584.383.884.0

      Mel spectrum + UAV images mAP@0.5：0.95

      91.2%

      		Label 193.293.793.4
      Label 294.392.993.6
      Label 395.692.494.0
      Label 494.995.195.0
      Label 595.595.795.6

      STFT+UAV images mAP@0.5：0.95

      96.6%

      		Label 1100.099.799.8
      Label 299.297.798.4
      Label 399.997.498.6
      Label 499.0100.099.5
      Label 5100.099.399.6

    • Table 4. Performance comparison between the proposed model and the traditional model

      View table

      Table 4. Performance comparison between the proposed model and the traditional model

      ModelPreprocessing time /sIdentification time /sTotal time /s
      RBF-EMD model1.140.51.64
      SVM model0.30.30.6
      Proposed model1×10-42×10-32.1×10-3
    Tools

    Get Citation

    Copy Citation Text

    Kang Xue, Kun Liu, Junfeng Jiang, Shuang Wang, Tianhua Xu, Zhenshi Sun, Sichen Li, Yuelang Huang, Xibo Jin, Tiegen Liu. Optical Fiber Vibration Sensing Detection with High Accuracy Based on YOLOv5s Model[J]. Acta Optica Sinica, 2023, 43(2): 0228001

    Download Citation

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

    Category: Remote Sensing and Sensors

    Received: Jun. 27, 2022

    Accepted: Aug. 1, 2022

    Published Online: Feb. 7, 2023

    The Author Email: Liu Kun (beiyangkl@tju.edu.cn)

    DOI:10.3788/AOS0228001

    Topics

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