A multi-peak detection algorithm for Fiber Bragg Grating sensing systems

• Published in: Optical fiber technology 2020-09, Vol.58 (C), p.102311, Article 102311
• Main Authors: Xu, Ouyang, Liu, Jingen, Tong, Xinglin, Zhang, Cui, Deng, Chengwei, Mao, Yan, Yin, Renjie, Jin, Chunjiao, Fang, Dingjiang
• Format: Article
• Language: English
• Subjects: Fiber Bragg Grating (FBG); Gaussian-LM algorithm; Hilbert transform; Peak-seeking algorithm; Reflection spectral
• ISSN: 1068-5200; 1095-9912
• DOI: 10.1016/j.yofte.2020.102311

•A multi-peak detection algorithm for FBG optical sensing systems is proposed.•A seven-point moving average filter is used to process the FBG reflection spectral signal.•Combining Hilbert transform with Gauss-LM algorithm for peak localization.•Analyze the effect of different segmentation thresholds on peak-seeking.•It has high accuracy compared with comparison algorithm. Aiming at the problem that traditional peak-seeking algorithms cannot directly detect multiple reflections of Fiber Bragg Grating (FBG) sensing systems, this paper proposes a multi-peak detection algorithm. A seven-point moving average filter is used to process the FBG reflection spectral signal, reduce the high-frequency noise in the spectral signal, and determine the optimal segmentation spectral threshold point through Hilbert transform. The derived spectral signal obtains multiple FBG reflection spectral signal’s sub-spectra and achieves the initial positioning of the spectral peak. The Levenberg-Marquardt (LM) algorithm is used to extract the Bragg wavelength from the segmented sub-spectral signal, as well as optimizing the Gaussian curve fitting coefficients. The optimized Gaussian-LM algorithm is then utilized to achieve precise positioning of the spectral peak. The influence of different spectral segmentation thresholds on FBG peak-seeking accuracy was analyzed. Theoretical analysis and experimental results showed that the proposed algorithm could dynamically detect multiple reflection spectra of the FBG sensing system with good stability while reducing the amount of peak-seeking data, which is highly beneficial to improve the signal demodulation rate. The average error of the algorithm is 4.34 pm, i.e., it has high accuracy compared to the comparison algorithm. As a result, it is shown that the multi-peak detection algorithm proposed in this paper can be applied to the distributed FBG sensing systems.