Research On Intrusion Vibration Signal Recognition Method Of Optical Fiber Perimeter Protection System

The fiber perimeter protection system identifies different intrusion behaviors by monitoring fiber vibration signals.The key to intrusion behavior identification is the feature characterization and extraction of fiber vibration signals.The main difficulty is that in the process of feature extraction,the decomposition of fiber vibration signal has the problems of weak adaptability and end effect.Therefore,how to use signal processing technology to adaptively decompose the fiber vibration signal and suppress the end effect is a subject are worth studying.This thesis focuses on the characteristics of fiber-optic vibration signals,and studies an improved overall empirical mode decomposition method and local mean decomposition method to extract fiber vibration signal characteristics and improves the recognition accuracy of the fiber perimeter protection system.The main work of this thesis is as follows:1.The domain characteristics of typical perimeter vibration signals are studied.After analyzing and comparing the advantages and disadvantages of different fiber optic sensors,an optical fiber perimeter guard system structure based on double Mach-Zade sensors,optical path mathematic model of double Mach-Zard sensor.The framing method of optical fiber signals was studied.2.An improved method of overall modal decomposition fiber signal processing is studied,and the white noise amplitude coefficient and the overall average number of times can be adaptively determined for different types of fiber-optic vibration signals to solve the problem of the total empirical mode decomposition method for the decomposition of vibration need to set parameters by human experience.The method of local mean decomposition is studied,and the fiber vibration signal is decomposed into several product functions to solve the end effect problem of fiber signal decomposition.3.A variety of vibration intrusion signal type recognition algorithms based on support vector machine classification is studied.For multiple types of fiber vibration signals,a multiple classifier based on binary tree support vector machine is designed.The fiber laser vibration signals are identified and simulated by using the method described in this thesis,and the recognition results are compared by using the methods of traditional total empirical mode decomposition and wavelet decomposition,the results show that the method described in this thesis can effectively identify the type of fiber vibration signal intrusion and improve the accuracy of system identification.