采用时域和复域小波变换的光纤周界振动信号识别

Optical Fiber Perimeter Vibration Signal Recognition Based on Time-domain and Complex Wavelet Transform

针对光纤周界安防系统存在报警率低和误报率高的问题,提出一种时域和复域小波变换相结合的光纤周界振动信号识别方法,以降低误报率、提高报警率并确定振动时间。该方法首先用光纤信号的时域小波特征来抑制干扰信号,排除静默信号并判断是否有振动发生;其次用信号复域小波特征中的能量分布和相位特征构造出光纤信号的识别特征;最后用支持向量机作为分类器训练光纤振动信号并识别出振动类型。

Aiming at fiber perimeter security system's low alarm and high false alarm rate,a method which combining the time-domain and complex wavelet transform to recognize the optical fiber perimeter's vibration signals was proposed so as to improve alarm rate and reduce false alarm rate as well as determine vibration time. In which,having the time-domain wavelet features of the optical fiber signal adopted to restrain the interference signals and exclude the silent signals as well as to judge whether the vibration occurred there; then having the energy distribution in signal's complex wavelet features and phase features employed to extract recognition features of the optic fiber signals,and finally having the support vector machine( SVM) taken as the classifier to train signals and identify the vibration types.