Perimeter Intrusion Defense System Based On MET Geophone

The detection sensors used in current perimeter intrusion defense system mainlyhave sound, vibration, passive infrared, magnetic, electrical, video sensor etc. Butmany physical quantities are easily influenced by external environment, theperformance of system based on those kinds of sensors are difficult to achieve higherstandards, seismic geophone has enormous advantage in doppler effect andenvironmental impact, therefore it is widely used in the perimeter intrusion defensesystem. But currently commonly used seismic geophones always have higher naturalfrequency and lower sensitivity, then their detection distance are shorter and they loseparts of low frequency component which target recognition needed, so it is necessaryto find a kind of ultra low frequency seismic geophone with high sensitivity. METgeophone has high sensitivity and ultra low natural frequency which can reach1Hz,so this paper develop the perimeter intrusion defense system based on MET geophone,at the same time, select most commonly used traditional geophone—JF-20DX-10Hzas a contrast detection sensor. The main contents of this paper are:Firstly, study the basic theory of the seismic signal, analyze the generation andpropagation mechanism of earthquake wave generated by target motion, elaboratedetail characteristics of rayleigh wave which contains the mainly energy of theseismic wave, do the theoretical foundation for the subsequent data processing.Secondly, introduce the structure and working principle of MET seismicgeophone, and design a series of experiments to test its actual performance,respectively, self noise test, sensitivity test, phase frequency characteristic test. Theself noise test use international standard in seismic field to test the noise accelerationpower spectral density and velocity power spectral density; the sensitivity and thephase frequency characteristic test use ultra low frequency vibration table byfrequency sweep to get the sensitivity and phase frequency characteristic. The abovethree experimental proof the actual performance of MET geophone conform itscalibration index.Thirdly, study the target recognition algorithm of wavelet packet combine BPneural network. Firstly use zero-crossing rate and kurtosis analysis in the time domainto analyze the seismic signal, use power spectrum analysis in the frequency domain to analyze signal. Secondly introduce the principle of wavelet analysis, and introduce thecommonly used wavelet functions, learn the principle of wavelet denoising algorithm,study the improved wavelet analysis algorithm——wavelet packet analysis and itsapplication of feature extraction. Then introduce the principle of BP neural network,learn the network structure, training process and its application in target recognition.Finally establish the target recognition algorithm model based on wavelet packetcombine BP neural network, establish the feature extraction algorithm of waveletpacket, establish the BP neural network and introduce its training process in detail.Fourthly, develop the perimeter intrusion defense system software based on NILabVIEW, program good man-machine interface to realize data display, storage andprocessing. Use NI TDMS file format for data preservation; use LabVIEW callMATLAB mode to realize target signal feature extraction and BP neural networktraining and recognition, the recognition results display directly on the main interfaceof the system software. In order to improve the accuracy of the target identification,increase the off-line analysis module for further analysis.External experimental results show that, compared with JF-20DX-10Hzgeophone, perimeter intrusion defense system based on MET geophone has higherrecognition accuracy and recognition distance.