Phase Sensitive Optical Time Domain Reflectometry Based on Compressive Sensing

A sub-Nyquist random single-pulse sampling method based compressive sensing technique is proposed firstly to reconstruct signals under test in phase-sensitive optical time domain reflectometry (φ-OTDR). The key information of Rayleigh backscattering curves can be recognized after sparse transformation, which can be successfully reconstructed with far smaller number of measurements by orthogonal matching pursuit algorithm. Unlike the uniform sampling in conventional φ-OTDR system, pulse intervals are randomly modulated to generate a random single-pulse sequence in the proposed method in order to achieve random single-pulse sampling along every sensing point of the fiber. Experimental results verify that pulse sequences with average pulse repetition rate less than 2-kHz can be efficiently utilized to reconstruct an 8-kHz signal under test and obtain its frequency information, as well as to realize the detection of high frequency signals at different locations along the sensing fiber. It is demonstrated that the sampling rate is reduced greatly without any hardware modification compared to that required in the conventional method. This method can break the limits of the traditional Shannon-Nyquist sampling theorem to reconstruct signal, which is of great significance for the sensor performance enhanced in traditional φ-OTDR system.