Signal-to-noise ratio enhancement of sub-sampling phase-sensitive optical time-domain reflectometry sensing through weighted moving average and moving differential algorithm

The response frequency of a uniform sampling phase-sensitive optical time-domain reflectometry vibration sensing system is constrained by the sensing distance, for which non-uniform sub-Nyquist sampling can be employed to extend the detectable frequency range. However, sub-sampling corresponds to a low sampling rate, and when conventional signal recognition algorithms such as moving average and moving differential (MA-MD) are used, the achieved signal-to-noise ratio (SNR) is often not satisfactory. To address this issue, we propose a weighted MA-MD algorithm, in which the signal with more energy can be amplified more than the noise. For additive random sampling, the effectiveness of the proposed algorithm is evaluated through experiments conducted at various frequencies over a 6-km sensing distance. Compared with the conventional MA-MD and image edge detection methods, a significant SNR enhancement was achieved, and 109-kHz vibration was successfully retrieved with maximum and average sampling rates of 16.67 and 4.04 kHz, respectively.