φ-OTDR signal compression scheme based on the compressed sensing theory

Abstract

In this paper, based on the compressed sensing theory and the orthogonal matching pursuit algorithm, we have designed a data compression scheme, taking the Space-Temporal graph, time domain curve, and its time-frequency spectrum of phase-sensitive optical time-domain reflectometer as the target signals. The compression rates of the three signals were 40%, 35%, and 20%, while the average reconstruction times were 0.74 s, 0.49 s, and 0.32 s. The reconstructed samples effectively retained the characteristic blocks, response pulses, and energy distribution that symbolize the presence of vibrations. The average correlation coefficients of the three kinds of reconstructed signals with the original samples were 0.88, 0.85, and 0.86, respectively, and then a series of quantitative metrics were designed to evaluate the reconstructing efficiency. We have utilized the neural network trained by the original data to identify the reconstructed samples with an accuracy of over 70%, indicating that the reconstructed samples accurately present the vibration characteristics.