Fast and Accurate Brillouin Optical Time-Domain Sensing by Sparse Frequency Sampling and ANN-Based Recover Method

We propose a novel sparse frequency sampling and Brillouin scattering spectrum (BSS) recover method to greatly reduce the acquisition time of the frequency-sweep-based Brillouin optical time-domain sensor (BOTDS) without increasing hardware complexity. In the proposed method, the acquired BSSs only contain a few frequency points and thus show low resolution due to the sparse frequency sampling strategy. Then, an artificial neural network (ANN) is used to accurately recover the BSSs with sparse frequency sampling to that with normal frequency sampling and thus achieve fast and accurate temperature/strain measurement. In a proof-of-concept experiment, we measured the BSS of a 3-km sensing fiber and achieved 3-m spatial resolution by using a Brillouin optical time-domain reflectometry (BOTDR) sensor. Through sparse frequency sampling, the data acquisition time is 5.5% with respect to that of the normal sampling. When the average (AVG) time of BSS is higher than 1000, the temperature measurement uncertainty is 0.2 degrees C.