Laser & Optoelectronics Progress, Volume. 62, Issue 17, 1706002(2025)

Intelligent Processing Technology for Anomaly Detection and Denoising of Optical Fiber Plasma Sensing Data

Binqiang Ye1, Yayu Yin1, Minglang Zhang1, Xiaoling Peng2、*, and Bin Tang2
Author Affiliations
  • 1School of Artificial Intelligence, Chongqing University of Technology, Chongqing 401135, China
  • 2School of Electrical and Electronic Engineering, Chongqing University of Technology, Chongqing 400054, China
  • show less
    Figures & Tables(17)
    Original data and synthesized data. (a) Original data; (b) synthesized data
    DBSCAN classification results for three groups of outlier regions. (a) Classification results of the first group of data; (b) classification results of the second group of data; (c) classification results of the third group of data
    Outlier detection results for synthesized data
    First decomposition results and IMF classification results. (a) First decomposition results; (b) IMF classification results
    Second decomposition results
    Denoising results for synthesized data
    Schematic diagram of optical fiber plasma sensor detection device
    Spectral data and spectral data at the 2000th time point. (a) Spectral data; (b) spectral data at the 2000th time points
    Time series of data at the spectral Dip and processed results. (a) Time series of data at the spectral Dip; (b) processed results of the time series
    Average intensity curve and noise level curve
    • Table 1. Statistical results of evaluation metrics for outlier detection algorithms

      View table

      Table 1. Statistical results of evaluation metrics for outlier detection algorithms

      ModelAccuracyPrecisionRecallF1 score
      Z-score0.98880.87960.90480.8920
      IQR0.98930.91090.87620.8932
      KNN0.98360.84880.82860.8386
      iForest0.98850.89760.87620.8867
      IRLS0.98800.99390.77140.8686
      AutoEncoder0.88260.96390.91080.9366
      LSTM-AD0.88700.96640.91310.9390
      RirPLS0.99880.98120.99520.9882
    • Table 2. AAPE and RPE for each IMF component

      View table

      Table 2. AAPE and RPE for each IMF component

      IMF componentAAPERPEIMF componentAAPERPE
      IMF10.92920.0035IMF70.18330.4709
      IMF20.69400.0318IMF80.16890.4829
      IMF30.46380.1363IMF90.15390.4956
      IMF40.32050.2741IMF100.15180.4925
      IMF50.25420.3762IMF110.06570.8734
      IMF60.20830.4394
    • Table 3. Mutual information between IMF and a priori data

      View table

      Table 3. Mutual information between IMF and a priori data

      IMF componentNorMIIMF componentNorMI
      IMF10.0031IMF70.0772
      IMF20.0046IMF80.1278
      IMF30.0084IMF90.1633
      IMF40.0171IMF100.2177
      IMF50.0271IMF110.3111
      IMF60.0421
    • Table 4. Comparison results of different denoising algorithms

      View table

      Table 4. Comparison results of different denoising algorithms

      ItemEMDCEEMDANICEEMDANProposed
      RMSE0.44010.39670.38020.3711
      SNR /dB16.572217.472417.843218.0539
    • Table 5. Evaluation indicators in multiple scenarios

      View table

      Table 5. Evaluation indicators in multiple scenarios

      Input SNR /dBOutlier countAccuracyPrecisionRecallF1-scoreRMSESNR /dB
      02100.99850.98570.98570.98570.820511.1619
      4200.99660.97650.99050.98350.907710.2844
      6300.99510.97620.99190.98400.95279.8641
      52100.99980.99531.00000.99760.490615.6292
      4200.99830.99280.99050.99170.592113.9947
      6300.99930.99521.00000.99760.611212.7197
      102100.99880.98120.99520.98820.371118.0539
      4200.99730.98280.99010.98650.399517.4129
      6300.99710.98250.99840.99040.399717.4083
    • Table 6. Repeatability analysis results in multiple scenarios

      View table

      Table 6. Repeatability analysis results in multiple scenarios

      Input SNR /dBOutlier countEvaluation metricsOriginal dataProcessed data meanProcessed data SD
      0210NI5.14970.66362.8×10-9
      CD1.55591.14651.6×10-11
      AAPE0.99380.61194.0×10-11
      420NI6.07920.57103.8×10-10
      CD1.60190.94173.5×10-9
      AAPE0.99380.52945.0×10-11
      630NI6.69180.54445.1×10-10
      CD1.65010.91443.0×10-10
      AAPE0.99300.49731.8×10-13
      5210NI2.85100.56505.4×10-9
      CD1.59700.90941.8×10-13
      AAPE0.99500.54322.5×10-13
      420NI4.11890.64144.3×10-8
      CD1.65060.87751.5×10-13
      AAPE0.99470.52423.2×10-9
      630NI5.28120.66952.2×10-11
      CD1.66710.91012.2×10-13
      AAPE0.99410.52194.4×10-13
      10210NI1.83000.52547.2×10-10
      CD1.64201.15268.0×10-12
      AAPE0.99730.61943.0×10-13
      420NI2.42580.54553.2×10-10
      CD1.65030.89359.2×10-11
      AAPE0.99580.52321.6×10-9
      630NI2.88960.53309.2×10-9
      CD1.66490.88345.1×10-11
      AAPE0.99560.51941.3×10-10
    • Table 7. Comparison of NI, CD, and AAPE values of original data and data processed by different algorithms

      View table

      Table 7. Comparison of NI, CD, and AAPE values of original data and data processed by different algorithms

      ItemOriginal dataEMDCEEMDANICEEMDANProposed
      NI1.01360.67880.53640.52720.5118
      CD0.92750.77670.76310.75890.7478
      AAPE0.99730.33440.31560.32620.3041
    Tools

    Get Citation

    Copy Citation Text

    Binqiang Ye, Yayu Yin, Minglang Zhang, Xiaoling Peng, Bin Tang. Intelligent Processing Technology for Anomaly Detection and Denoising of Optical Fiber Plasma Sensing Data[J]. Laser & Optoelectronics Progress, 2025, 62(17): 1706002

    Download Citation

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

    Category: Fiber Optics and Optical Communications

    Received: Apr. 9, 2025

    Accepted: May. 7, 2025

    Published Online: Sep. 16, 2025

    The Author Email: Xiaoling Peng (Pengxiaoling@cqut.edu.cn)

    DOI:10.3788/LOP251129

    CSTR:32186.14.LOP251129

    Topics