基于多特征参量的?-OTDR分布式光纤扰动传感系统模式识别研究

Study of Pattern Recognition Based on Multi-Characteristic Parameters for ?-OTDR Distributed Optical Fiber Sensing System

针对目前相位敏感光时域反射计(?-OTDR)分布式光纤扰动传感系统难以对扰动事件进行有效判别的问题,提出一种基于多特征参量的扰动模式识别方法。作为识别依据的多特征参量包括平方差、短时过电平率、短时傅里叶变换以及扰动持续时间,对以上特征参量利用加权平均型综合评判函数进行模式识别,可以有效识别出应力破坏(敲击)、攀爬、浇水(模拟下雨环境)以及轻度碾压(模拟非破坏性人为扰动)四种扰动模式。对60组共240个实验样本进行识别,对以上四种扰动事件的识别率分别达到91.2%,88.3%,90%和86.7%。该方法克服了现有单一特征参量识别模式受限的问题,提高了扰动事件识别的准确性。

In view of the present phase sensitive optical time-domain reflectometer （Ф-OTDR） distributed optical fiber sensing system is difficult to distinguish the disturbance effectively. A disturbance modes recognition method based on multi-parameters is presented. A weighted average synthetic evaluation function is used for characteristic parameters pattern recognization including square difference, short time level crossing rate, short time Fourier transform and duration of the disturbance. By the proposed method, four disturbance modes including stress damage （breaking）, climbing, watering （simulate rain environment） and mild crushing （simulate harmless artificial disturbance） are effectively recognized. The recognition rate of disturbance pattern recognization is realized as 91.2%, 88.3%, 90% and 86.7%, respectively, for the corresponding four disturbance modes by analyzing the statistics of 60 groups of 240 samples. The proposed method breaks though the limitation of disturbance modes recognized by current single-characteristic parameter method and improves the reliability of disturbance pattern recognization.