Location Method Based on Support Vector Machine for Distributed Sagnac Fiber Sensing System

In order to simply locate the pipeline leakage, a location method based on support vector machine (SVM) for distributed Sagnac fiber sensing system is proposed. The sensing fiber is segmented according to the required resolution, and the location of different fiber segments is converted into a multi-classification problem of interference signals caused by external disturbances. The signal feature is extracted by spectral transform, and classified by a SVM classifier. Simulation results show that the classification rate can up to 96.61% and 100% at the resolutions of 50 m and 100 m, respectively. The method is simple and effective, insensitive to noise, and the location resolution is adjustable.