Research On Pattern Recognition Of Grating Array Perimeter Intrusion Alarm System

Extensive research has been carried out based on the needs of optical fiber sensing systems in the fields of vibration measurement,temperature measurement,strain,and acoustic detection.As a new generation of optical fiber sensing technology,grating array sensing technology solves the technical limitations of traditional optical fiber or a single grating as the sensing unit.Applying the grating array technology to the perimeter security system,its advantages such as high signal-to-noise ratio,wide monitoring range,and simple installation gradually occupy a place in the perimeter field.This article starts with the evolution of the perimeter system and analyzes and compares the perimeter system based on the grating array technology.The research and exploration of pattern recognition are carried out on the false alarm problem of the intrusion system of the grating array perimeter.The main research contents of this thesis are as follows:(1)The system architecture of the grating array perimeter and the basic principle of using grating array technology to realize vibration sensing are described,and the application of grating array fiber technology in the security field is analyzed and summarized.(2)By building a grating array perimeter system for data collection,the data characteristics of the grating array perimeter system are analyzed from the data level,and finally the signal characteristics are preprocessed.(3)A perimeter intrusion recognition method based on grating array fiber wavelet neural network is proposed.Specifically,combining the signal characteristics of the wavelet neural network and the grating array grating perimeter sensing system,a wavelet neural network model suitable for the perimeter system is constructed,and the simulation verifies the effectiveness of the model's recognition.Aiming at the application of grating array perimeter security in different scenarios,a pattern recognition library is designed and established,which is convenient for flexible use in engineering practice,adapts to local conditions,and promotes the intelligentization of optical fiber security.(4)The proposed method for identifying the perimeter intrusion pattern of wavelet neural network based on grating array is validated according to the standards of engineering implementation.The experimental results show that for the four typical patterns: walking,climbing,and dragging.The overall recognition rate of shaking reached 90%.It is of reference significance for the new generation of grating array perimeter system to solve and optimize false alarms.