A Novel Method of Recognizing Disturbance Events in Φ-OTDR Based on Affinity Propagation Clustering and Perturbation Signal Selection
Shaohua Xu, Zujun Qin, Wentao Zhang, Xianming Xiong, Heng Li, Zexian Wei, Octavian Postolache, Chao Mi
Abstract
In order to reduce computational redundancy, and to recognize the perturbation signals quickly and efficiently, this paper proposed a novel recognition method for Φ-OTDR system based on affinity propagation clustering. Variational mode decomposition (VMD) are utilized to accurately extract the perturbation signal characteristic information buried in the vibration signals. Moreover, in order to improve the efficiency of clustering algorithm, a novel perturbation signal selection technique was developed in this work, whose effectiveness was experimentally verified. The proposed novel recognition method is then applied to the perturbation signal detection in Φ-OTDR system. The obtained results demonstrated that the proposed method is reliable and accurately identify different perturbation categories of signals.