The Multi Target Recognition Of Intrusion Signal Of Perimeter Security With Distributed Fiber-optic Sensor

SPS(Security&Protection System consists series of security measures andproducts involving video monitoring system, infrared alarm system, riot securitysystem, which is based on supporting public safety. While perimeter security systemof distributed fiber-optic is a advanced SPS specially used in National security field,which based on modern photoelectric technology and information processingtechnology, and also use optical fiber vibration sensor as its core-block. As thedistributed optical fiber sensing technology is high sensitive, low energy, strongrobustness, long detection distance, it has super value in border where the terrain iscomplex, the environment is very poor and the people there are scarce. So thedistributed optical fiber sensing technology is a research emphasis of border securitysystems all over the world.In this paper, a new method to distinguish the environment noise and theinvasion, the different kinds of invasions is proposed by first processing the originalsignal with wavelet analysis and wavelet threshold denoising, then recognizing themulti-invasion based on BP neural network.The main important work of this paper:(1) analysis the value and components of perimeter security system; introduce theworking principle of fiber optical sensor, the model and technical features ofdistributed fiber sensor and interferential fiber optic sensor; compare somedifferent kinds of interferential fiber optic sensors, finally choose the M-Zinterferometers as the basic model of this paper.(2) first the eigenvectors of original signals is extracted by using the signalprocessing system of distributed fiber perimeter security system, mathematicalderivation process based on wavelet analysis and wavelet threshold de-noising,which explodes the original signal into7pieces, then draw the energy spectrumof each signal piece, analysis the difference between the eigenvectors of invasionsignal and interfering signal, finally propose the method to deal with the originalsignal.(3) The recognition function of the system is based on BP neural network. Weoptimize the algorithm and increase the recognition rate of recognition functionof the system. In the perimeter security system of distributed fiber-optic, theoriginal invasion signal and interfering signal will be all received by receiving end. The feature vectors of every signal received by receiving end will berecognized effectively by the signal processing system, so that we candistinguish which is useful. We should put this feature vectors into recognitionfunction and set the threshold value to recognize the signal.(4) first the eigenvectors of original signals is extracted by using the signalprocessing system of distributed fiber perimeter security system, mathematicalderivation process based on wavelet analysis and wavelet threshold de-noising,which explodes the original signal into7pieces, then draw the energy spectrumof each signal piece, analysis the difference between the eigenvectors of invasionsignal and interfering signal, finally propose the method to deal with the originalsignal.