Multi-peak detection algorithm for FBG array sensing based on self-adaptive thresholding

When dealing with the reflected signals from an array of optical fiber Bragg gratings (FBGs) in fiber optic sensing, the conventional multi-peak detection algorithm often faces challenges due to the presence of noise interference, potentially resulting in demodulation failures. In this study, we propose a robust self-adaptive multi-peak detection algorithm. First, the reflected signals from the optical fiber Bragg grating array are normalized to enhance the stability of the algorithm. Next, an improved thresholding function in a wavelet transform denoising method is introduced to process the normalized FBG signals, effectively reducing high-frequency noise within the signals. Following this, the spectrum is segmented using the Hilbert transform and a self-adaptive threshold mathematical model, and then achieve stable 3 dB bandwidth spectrum segmentation by using spectrum expansion techniques. Lastly, the traditional peak detection algorithm is applied to extract the Bragg wavelengths from the segmented sub-spectral signals. Theoretical analysis and experimental results provide comprehensive evidence that employing a self-adaptive threshold for spectral segmentation significantly enhances the algorithm's portability across diverse scenarios, thereby improving the demodulation speed and stability of the algorithm. The proposed algorithm provides a precise and noise-resistant demodulation method for handling multi-peak signals in quasi-distributed sensing networks.