SNR-Based Denoising Dynamic Statistical Threshold Detection of FBG Spectral Peaks

This paper targets a Denoising Dynamic Statistical Threshold (DDST) detection algorithm to detect the “presence” of Fiber Bragg Grating (FBG) spectral peaks in noise with changing Signal-to-Noise Ratio (SNR) sensing channel. Computing DDST is based on statistical parameters background noise. The determined by adjusting it using SNR via determining targeted probability false alarms ( p FA ). proposed incorporates effect fluctuations, nonlinear signal attenuation Single-Mode (SMF), as well influence short-term interference. implemented sliding wavelength window technique conjunction FBG power scaling allow automatic and DDST. During possible FBGs resonant wavelengths overlap resulting from approaching/colliding responses FBGs, also improves robustness resolving these overlaps. marginally takes into account shapes peaks. Advantageously, independent shapes. Our simple implement. Measurements done two Optical Spectral Analyzers (OSAs) confirmed significant improvements reduction (i.e. denoising), noisy smoothing SNR, improved adjacent detectability resolving. experiments usability under severe network conditions (with low reflected below $-$ 75 dB xmlns:xlink="http://www.w3.org/1999/xlink">SNR notation="LaTeX">$< $ 4 standard deviation notation="LaTeX">$\sigma >$ 7 fluctuations) resolution 3.43 pm.