Abstract
One of the important successes of optical fiber sensor established for the security system is the detection and the recognition of any type of events. The performance parameters (event recognition, event detection position, and time of detection) are unavoidable and describe the validity of any perimeter detection system. An event recognition is any signal detected within the protected area, and it is related to a non-intrusion event and an intrusion event. To achieve the detection and the recognition events at the real time, an effective two-level vibration recognition method and a technique are proposed and presented in this article. The signal characteristics (short-term energy and short-time over-threshold) have been used and compared to the dynamic threshold to judge the type of event. Then the extraction of the power distribution features on the frequency domain through power spectral estimation on the suspected intrusion signal samples is carried out and finally combined with the time-domain characteristics as feature vector through Support Vector Machine to determine the efficiency and effectiveness of the proposed vibration recognition method. The experimental simulation results show that the proposed method is effective and reliable. With collected data, it can detect and recognize the type of event in real time.
Highlights
With a broad social development and a scientific and technological progress, the distributed perimeter security system with fiber as sensor has been applied to oil and gas pipelines, prisons, military bases, and so on, because of its high sensitivity, good safety performance, timely response, wide monitoring range, and simple installation, where the field monitored by the current perimeter security system looks for hotspots.The basic indicators of the perimeter fiber safety system are the system identification accuracy and the detection response time
The accuracy of system identification is related to the signal pattern recognition algorithm and the key is the selection of signal characteristics, while the detection response time depends on the complexity of the algorithm
The short-term Fourier transform is used to qualitatively analyze the spectral features that are in Liang et al.[5] and Li et al.,[7] but in DiCarlo and Weber[4] and Zhu et al.,[6] the energy distribution and vibration duration of different frequency bands are extracted in the complex wavelet domain as signal characteristics, and the Support Vector Machine (SVM) is used to classify them
Summary
With a broad social development and a scientific and technological progress, the distributed perimeter security system with fiber as sensor has been applied to oil and gas pipelines, prisons, military bases, and so on, because of its high sensitivity, good safety performance, timely response, wide monitoring range, and simple installation, where the field monitored by the current perimeter security system looks for hotspots. The short-term Fourier transform is used to qualitatively analyze the spectral features that are in Liang et al.[5] and Li et al.,[7] but in DiCarlo and Weber[4] and Zhu et al.,[6] the energy distribution and vibration duration of different frequency bands are extracted in the complex wavelet domain as signal characteristics, and the Support Vector Machine (SVM) is used to classify them. In view of the system’s need for effectiveness and real time, two-level vibration recognition method is proposed in this article and the steps are as follows: first, preliminary discrimination of event types is performed by time-domain features, and the nonsuspected disturbance data are directly produced as a non-intrusion type. The article focuses on the characteristics of modern power spectrum of different event types, to verify and analyze the accuracy and real time of system identification.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
