论文部分内容阅读
相位敏感光时域反射计(Ф-OTDR)在光纤围栏等动态传感领域具有重要的应用,快速、有效地对入侵信号分类识别有着十分重要的意义。基于频谱分析提出了一种称为频谱欧氏距离法(EDFS)的快速模式识别方法。该方法通过短时平移差分和短时能量法对Ф-OTDR的解调信号进行提取,确定待分析数据段;对数据段进行归一化和快速傅里叶变换,获得信号的频谱特征;计算信号频谱与预先生成的模板之间的欧氏距离对入侵信号进行分类、识别。采用三种入侵信号对该方法的有效性和实时性进行了实验验证。结果表明,该模式识别方法可以有效识别扰动信号,识别时间小于传统的动态时域规划模式识别方法耗时的1/10。同时,该方法所需训练样本较少,对环境噪声有一定程度的抑制作用。
Phase-sensitive optical time-domain reflectometer (Ф-OTDR) is an important application in the field of dynamic sensing such as fiber optic fence. It is very important for classification and recognition of intrusion signals quickly and effectively. Based on spectrum analysis, a fast pattern recognition method called spectral Euclidean distance method (EDFS) is proposed. In this method, the demodulated signal of Ф-OTDR is extracted by short-time translation difference and short-time energy method to determine the data segment to be analyzed; the data segment is normalized and fast Fourier transformed to obtain the spectral characteristic of the signal; The Euclidean distance between the signal spectrum and the pre-generated template classifies and identifies the intrusion signal. Three kinds of intrusion signals are used to verify the effectiveness and real-time performance of the proposed method. The results show that the pattern recognition method can effectively identify the disturbance signal, and the recognition time is less than 1/10 of the traditional dynamic time-domain pattern recognition method. At the same time, this method requires fewer training samples and has a certain degree of inhibition of environmental noise.