基于联合特征的两级光纤入侵信号识别方法
|
摘要
针对Sagnac光纤传感系统在识别过程中存在振动信号提取不完善,特征向量表征振动信号不充分,导致对入侵信号识别率不高的问题,提出一种基于联合特征的两级识别算法。该算法第一级提出融合短时过电平和短时对数能量的LC(Level Crossing,LC)算法对光纤振动号提取;第二级采用小波包分解提取信号各频带能量与峭度特征,联合原信号和差分信号的时域特征,构成包含频域、时域、差分时域的振动信号联合特征,然后,构建支持向量机(Supported Vector Machine,SVM)识别入侵信号类型。实验结果表明:本文方法能有效提取振动信号,对入侵信号的识别率可达95. 4%。
A two stage discriminative method based on joint eigenvectors is proposed to settle the problem of inaccurate extraction and insufficient eigenvectors representation of vibration signal that results in low recognition rate in Sagnac vibration sensing system.in the first level,A level crossings algorithm that combine short-term level cross and short-term logarithmic energy is used for extracting vibration signal. In the second level,using wavelet packet decomposition to extract the energy and kurtosis characteristics of each frequency band,combining with the time-domain characteristics of the original signal and the differential signal. Form joint eigenvectors of vibration signal including frequency domain,time domain and differential time domain. Lastly,use Support Vector Machine as the classifier to identify intrusion type. The experiment results show that this method can extract vibration signal effectively. The recognition rate of the intrusion signal is 95. 4%.
A two stage discriminative method based on joint eigenvectors is proposed to settle the problem of inaccurate extraction and insufficient eigenvectors representation of vibration signal that results in low recognition rate in Sagnac vibration sensing system.in the first level,A level crossings algorithm that combine short-term level cross and short-term logarithmic energy is used for extracting vibration signal. In the second level,using wavelet packet decomposition to extract the energy and kurtosis characteristics of each frequency band,combining with the time-domain characteristics of the original signal and the differential signal. Form joint eigenvectors of vibration signal including frequency domain,time domain and differential time domain. Lastly,use Support Vector Machine as the classifier to identify intrusion type. The experiment results show that this method can extract vibration signal effectively. The recognition rate of the intrusion signal is 95. 4%.
引文
[1] HUANG X,WANG Y,LIU K,et al. Event Discrimination of Fiber Disturbance Based on Filter Bank in DMZI Sensing System[J]. IEEE Photonics Journal,2016,8(3):1-14.
[2] LIU K,TIAN M,LIU T,et al. A High-Efficiency Multiple Events Discrimination Method in Optical Fiber Perimeter Security System[J]. Journal of Lightwave Technology,2015,33(23):4885-4890.
[3]朱程辉,瞿永中,王建平.基于时频特征的光纤周界振动信号识别[J].光电工程.2014,1:16-22.
[4]刘琨,何畅,刘铁根,等.一种用于光纤周界安防系统的端点检测方法[J].光电子·激光.2014,11:2136-2140.
[5] KATSIFOLIS J. Robust event classification for a fiber optic perimeter intrusion detection system using level crossing features and artificial neural networks[C]. Orlando:Fiber Optic Sensors and Applications VI,2010:801-812
[6]吴红艳,贾波,卞庞.光纤周界安防系统端点检测技术的研究[J].仪器仪表学报.2013,4:743-748.
[7]蒋立辉,盖井艳,王维波,等.基于总体平均经验模态分解的光纤周界预警系统模式识别方法[J].光学学报.2015,10:60-66.
[8]喻骁芒,罗光明,朱珍民,等.分布式光纤传感器周界安防入侵信号的多目标识别[J].光电工程.2014,1:36-41.
[9]杨正理,孙书芳.基于小波能量熵的光纤周界安防系统信号识别[J].光电子·激光.2016,12:1328-1333.
[10]张春熹,李勤,李立京,等.基于Sagnac组合型干涉仪的光纤分布式扰动传感器(英文)[J].红外与激光工程.2014,10:3378-3382.
[11] KATSIFOLIS J. Elimination of rain-induced nuisance alarms in distributed fiber optic perimeter intrusion detection systems[C]. Orlando:Fiber Optic Sensors and Applications VI,2009:401-411
[12] SEEDAHMED,MAHMOUD,YUVARAJA,et al. RealTime Distributed Fiber Optic Sensor for Security Systems:Performance,Event Classification and Nuisance Mitigation[J].光子传感器(英文版),2012,2(3):225-236.
[13]张俊楠,娄淑琴,梁生.基于SVM算法的φ-OTDR分布式光纤扰动传感系统模式识别研究[J].红外与激光工程,2017,4:212-218.
[14] WANG Y,XIANG J,MARKERT R,et al. Spectral kurtosis for fault detection,diagnosis and prognostics of rotating machines:A review with applications[J]. Mechanical Systems and Signal Processing,2016,66–67:679-698.
[15] BIAN Y J,HUANG H M,WANG T T. A Research on the SVM Classification of Earthquake and Explosion Based upon Seismic Wave Features[J]. International Journal of Digital Content Technology&Its Applic,2014,26(5):888-898.
[2] LIU K,TIAN M,LIU T,et al. A High-Efficiency Multiple Events Discrimination Method in Optical Fiber Perimeter Security System[J]. Journal of Lightwave Technology,2015,33(23):4885-4890.
[3]朱程辉,瞿永中,王建平.基于时频特征的光纤周界振动信号识别[J].光电工程.2014,1:16-22.
[4]刘琨,何畅,刘铁根,等.一种用于光纤周界安防系统的端点检测方法[J].光电子·激光.2014,11:2136-2140.
[5] KATSIFOLIS J. Robust event classification for a fiber optic perimeter intrusion detection system using level crossing features and artificial neural networks[C]. Orlando:Fiber Optic Sensors and Applications VI,2010:801-812
[6]吴红艳,贾波,卞庞.光纤周界安防系统端点检测技术的研究[J].仪器仪表学报.2013,4:743-748.
[7]蒋立辉,盖井艳,王维波,等.基于总体平均经验模态分解的光纤周界预警系统模式识别方法[J].光学学报.2015,10:60-66.
[8]喻骁芒,罗光明,朱珍民,等.分布式光纤传感器周界安防入侵信号的多目标识别[J].光电工程.2014,1:36-41.
[9]杨正理,孙书芳.基于小波能量熵的光纤周界安防系统信号识别[J].光电子·激光.2016,12:1328-1333.
[10]张春熹,李勤,李立京,等.基于Sagnac组合型干涉仪的光纤分布式扰动传感器(英文)[J].红外与激光工程.2014,10:3378-3382.
[11] KATSIFOLIS J. Elimination of rain-induced nuisance alarms in distributed fiber optic perimeter intrusion detection systems[C]. Orlando:Fiber Optic Sensors and Applications VI,2009:401-411
[12] SEEDAHMED,MAHMOUD,YUVARAJA,et al. RealTime Distributed Fiber Optic Sensor for Security Systems:Performance,Event Classification and Nuisance Mitigation[J].光子传感器(英文版),2012,2(3):225-236.
[13]张俊楠,娄淑琴,梁生.基于SVM算法的φ-OTDR分布式光纤扰动传感系统模式识别研究[J].红外与激光工程,2017,4:212-218.
[14] WANG Y,XIANG J,MARKERT R,et al. Spectral kurtosis for fault detection,diagnosis and prognostics of rotating machines:A review with applications[J]. Mechanical Systems and Signal Processing,2016,66–67:679-698.
[15] BIAN Y J,HUANG H M,WANG T T. A Research on the SVM Classification of Earthquake and Explosion Based upon Seismic Wave Features[J]. International Journal of Digital Content Technology&Its Applic,2014,26(5):888-898.