基于互信息截取的光纤周界振动信号识别

摘要

Aiming at both strong nonlinearity and non-stationarity of vibration signals collected from optical fiber perimeter system, the mutual information method was applied to get delay time and then take it as signal’s maximal intercept length;and making use of complex wavelet packet transform to extract fiber-optic signal’s energy distribu-tion characteristics was implemented, including having support vector machine (SVM) adopted to identify types of intrusion vibration signals. Experimental results show that, this method can effectively recognize the signals invaded and improve the detection probability of the system as well as reduce the amount of calculation.%针对光纤周界系统采集的入侵振动信号存在的强非线性和非平稳性问题，以互信息法求取的延迟时间点数作为信号最大截取长度，采用复小波包变换提取光纤信号的能量分布特征，采用支持向量机作为分类器对入侵振动信号类型进行识别。实验结果表明：该方法可有效识别入侵信号，提高系统检测率，减小系统计算量。