Multiple disturbance detection and intrusion recognition in distributed acoustic sensing

摘要

Distributed acoustic sensing system can be used in the long-distance and strong-EMI condition for monitoring and inspection. In this paper, location method for optical fiber multiple dynamic disturbances signals is proposed to solve the difficulty with Distributed acoustic sensing (DAS) system in effectively locates multiple dynamic disturbances. The first step: locate multiple dynamic disturbances signals exactly by using the multiple threshold method. The second step: the Empirical Mode Decomposition(EMD) method and the Fourier transform(FFT) is proposed to extract the signal features. By analyzing the time domain signals of the intrusion location that we can look for the most efficient signal feature to form a pattern feature vectors for classification. After the first two steps, we can get feature vectors of different types of dynamic disturbances. By utilizing support vector machine(S VM) classifiers to identify feature vectors, patterns of intrusion events are recognized accurately. Experiments show that after using this method to process 300 dynamic disturbances samples generated by three different intrusion events, namely, passing, hurling and knocking, the location accuracy is about 1.6m, the recognition rates of intrusion events are over 90%.