Identification Of Intrusion Signals In Perimeter Security Based On FBG

With the rapid development of science and technology,the continuous progress of society and the improvement of people's security awareness,necessary measures should be taken to ensure the safety of perimeter.in important areas such as military sites,railways,factories and banks.However,with the advent of the information age,new criminal means supported by the internet,computers and electronic technologies are increasingly prominent,and criminal tactics tend to be intelligent,making traditional security technologies exposed more and more hidden dangers.In recent years,with the rapid development of optical fiber sensing technology,fiber optic perimeter system based on fiber sensing technology has become a research hotspot in perimeter security field.In particular,the security system based on Fiber Bragg Grating(FBG)sensor technology.This article has done in-depth research on new methods and key technologies based on FBG perimeter security,in order to improve the target recognition rate of FBG perimeter security system,and reduce false alarm rate and error rate.A series of research work has been carried out as follows:(1)A general method model for applying FBG to the perimeter security system was proposed.Combined with the overall requirements of the perimeter security system for the perimeter system,the research content based on the FBG signal analysis technology and pattern recognition method was established.The feasibility of implementing this method was demonstrated in several aspects including intellisense,weatherability,and sensitivity,etc.(2)According to the FBG sensing principle,especially the relationship between FBG wavelength and temperature and strain,a mathematical model based on FBG vibration sensing is established,and relevant theoretical analysis and calculations are performed.Several commonly used fiber grating demodulation methods are analyzed from the aspects of demodulation principle,practicality,etc.The tunable F-P filter method is determined as the demodulation method in this paper.(3)Based on the wavelet analysis theory,a new wavelet threshold denoising method is proposed.This method quickly decomposes the time domain alarm signal,and quantizes the wavelet coefficients through an appropriate threshold,effectively eliminating high frequencies caused by external vibration disturbances,makes the alarm accuracy rate further guaranteed.(4)A perimeter security intrusion recognition algorithm based on improved grey wolf optimization support vector machine was proposed.By analyzing the learning accuracy and learning ability of multi-samples of support vector machine,the optimization strategy of RBF kernel function was established and using the gray wolf algorithm to dynamically update the penalty factor and Gaussian nuclear width,combine the diversity features of particle swarm optimization algorithm to identify intrusion signals.(5)The experimental algorithm of the proposed algorithm is verified by the built-in FBG perimeter intrusion security system.Firstly,the learning performance of the PSOGWO-SVM algorithm proposed in this paper is tested.Compared with other group intelligent optimization algorithms,it verifies that the algorithm has faster convergence.Secondly,the intrusion signal is tested in the field.The experimental results show that the correct discriminative rate of this method is above 98%,the missed and false alarm rate is below 2%,and the correct pattern recognition rate is over 97%,achieve better recognition results.