Research on Vibration Event's Feature Extraction Method of Φ-OTDR System

Abstract

To be able to more accurately extract the characteristics of different vibration events in the Φ-OTDR system to better identify different vibration events on the optical fibre, this paper proposes a model method that combines variational modal decomposition and multi-scale permutation entropy. The vibration characteristics on the optical fibre are extracted, and the support vector machine is used for event classification, namely the VMD-MPE-SVM model. First of all, this paper performs a variational modal decomposition of the vibration signals collected by the Φ-OTDR system to obtain the eigenmode functions containing the characteristic information of different vibration events. The Second step is that using multiscale permutation entropy to quantify the vibration of each eigenmode function Characters. They are lastly inputting the high-dimensional feature vectors into the support vector machine's classifier to recognise different events. In the experiment, vibrations of different frequencies are performed on the optical fibre, and the recognition accuracy can reach 97.3%. Experimental results show that this model has good real-time performance and accuracy.