Acta Optica Sinica (Online), Volume. 2, Issue 13, 1305003(2025)

Image Denoising Method for BOTDR System Based on Adaptive Filtering and Non-Local Means

Jianyin Zhang1,2, Yuming Chen2,3, Yongzheng Li2,4, Linfeng Guo1,2,5、*, and Xiaomin Xu5,6,7
Author Affiliations
  • 1School of Physics and Optoelectronic Engineering, Nanjing University of Information Science & Technology, Nanjing 210044, Jiangsu , China
  • 2Jiangsu Provincial Engineering Research Center for Atmospheric Photoelectric Intelligent Sensing Technology, Nanjing 210044, Jiangsu , China
  • 3Jiangsu Product Quality Testing & Inspection Institute, Nanjing 210007, Jiangsu , China
  • 4China Railway No. 3 Group East China Construction Co., Ltd., Nanjing 211153, Jiangsu , China
  • 5Jiangsu International Joint Laboratory on Meterological Photonics and Optoelectronic Detection, Nanjing 210044, Jiangsu , China
  • 6Department of Engineering, University of Manchester, Manchester M13 9PL, United Kingdom
  • 7Department of Engineering, University of Cambridge, Cambridge CB2 1PZ, United Kingdom
  • show less
    Figures & Tables(17)
    Principle diagram of BOTDR
    BGS at room temperature and local temperature variation. (a) BGS three-dimensional diagram at room temperature; (b) schematic diagram of BGS with end temperature rise
    Schematic diagram of search window and similar window principle for NLM filtering algorithm
    Schematic diagrams of vertical gradient of BGS. (a) Three-dimensional schematic diagram of vertical gradient of BGS; (b) two-dimensional schematic diagram after taking the absolute value
    Noise reduction effects of three algorithms when the noise standard deviation is 5 dB. (a) Front view of 3D BGS diagram before noise reduction; (b) noise reduction effect of NLM algorithm; (c) noise reduction effect of BM3D algorithm; (d) noise reduction effect of GNLM algorithm
    Noise reduction effects of three algorithms when the noise standard deviation is 10 dB. (a) Front view of 3D BGS diagram before noise reduction; (b) noise reduction effect of NLM algorithm; (c) noise reduction effect of BM3D algorithm; (d) noise reduction effect of GNLM algorithm
    Noise reduction effects of three algorithms when the noise standard deviation is 15 dB. (a) Front view of 3D BGS diagram before noise reduction; (b) noise reduction effect of NLM algorithm; (c) noise reduction effect of BM3D algorithm; (d) noise reduction effect of GNLM algorithm
    Noise reduction effects of three algorithms when the noise standard deviation is 20 dB. (a) Front view of 3D BGS diagram before noise reduction; (b) noise reduction effect of NLM algorithm; (c) noise reduction effect of BM3D algorithm; (d) noise reduction effect of GNLM algorithm
    Schematic diagram of BODTR system device
    Schematic diagram of experimental design
    Schematic diagrams of two-dimensional BGS with noise reduction after heating at 50 ℃ with a pulse width of 50 ns. (a) Schematic diagram of two-dimensional BGS without noise reduction; (b) noise reduction effect of NLM algorithm; (c) noise reduction effect of BM3D algorithm; (d) noise reduction effect of GNLM algorithm
    BFS curves under the condition of a pulse width of 50 ns and heating at 30 ℃. (a) BFS curve without noise reduction; (b) BFS curve after NLM denoising; (c) BFS curve after BM3D denoising; (d) BFS curve after GNLM denoising
    Comparison of BFS curves of the front and the rear heating sections under the condition of a pulse width of 50 ns and heating at 30 ℃. (a) Comparison of noise reduction results of three algorithms in the front heating section; (b) comparison of noise reduction results of three algorithms in the rear heating section
    Comparison of BFS curves for the 50 °C heating section at the fiber rear end under different pulse widths. (a) Comparison of BFS without noise reduction; (b) comparison of BFS after NLM denoising; (c) comparison of BFS after BM3D denoising; (d) comparison of BFS after GNLM denoising
    • Table 1. Signal-to-noise ratio of the images, BFS uncertainty, and running time after noise reduction by three algorithms when the standard deviation of noise is 15 dB

      View table
      View in Article

      Table 1. Signal-to-noise ratio of the images, BFS uncertainty, and running time after noise reduction by three algorithms when the standard deviation of noise is 15 dB

      AlgorithmSignal-to-noise ratio /dBBFS uncertainty /MHzRunning time /s
      NLM33.622.1333
      BM3D43.071.2165
      GNLM42.181.2236
    • Table 2. Analysis of the influence of three algorithms on temperature accuracy after denoising

      View table
      View in Article

      Table 2. Analysis of the influence of three algorithms on temperature accuracy after denoising

      AlgorithmTemperature accuracy /%
      30 ℃ @ front end30 ℃ @ rear end50 ℃ @ front end50 ℃ @ rear end
      NLM5.1713.467.5118.31
      BM3D3.574.893.955.14
      GNLM3.352.012.963.02
    • Table 3. Evaluation of the influence of three algorithms on spatial resolution

      View table
      View in Article

      Table 3. Evaluation of the influence of three algorithms on spatial resolution

      AlgorithmEvaluation parameterLocal change intensity
      30 ns50 ns70 ns90 ns
      NLM

      Global

      Tenengrad

      0.08420.09130.11070.1137
      BM3D0.13410.13610.14430.1478
      GNLM0.13250.13740.14750.1520
    Tools

    Get Citation

    Copy Citation Text

    Jianyin Zhang, Yuming Chen, Yongzheng Li, Linfeng Guo, Xiaomin Xu. Image Denoising Method for BOTDR System Based on Adaptive Filtering and Non-Local Means[J]. Acta Optica Sinica (Online), 2025, 2(13): 1305003

    Download Citation

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

    Category: Optical Information Acquisition, Display, and Processing

    Received: May. 9, 2025

    Accepted: May. 23, 2025

    Published Online: Jul. 2, 2025

    The Author Email: Linfeng Guo (guolf_nj@163.com)

    DOI:10.3788/AOSOL250461

    CSTR:32394.14.AOSOL250461

    Topics