摘要
基于分布式光纤传感的地埋式入侵检测系统由于其无源、传输距离远、抗电磁干扰等特征,引起了人们的关注,并得到初步应用。系统中光纤传感信息的正确提取和事件识别是系统的关键性指标之一。文中提出了一种入侵信号的提取与识别算法。使用基于Hilbert变换的信号包络线提取,结合数学形态学上的腐蚀和膨胀运算思想,实现入侵信号片段的分割;采用小波阈值收缩方法,结合平移时不变算法实现入侵信号片段除噪;使用基于"小波包-能量"的方法提取信号特征;分别采用基于BP和RBF的人工神经网络实现对光纤信号分类。仿真实验验证了算法的有效性。
The perimeter intrusion detection system based on distributed fiber sensing has paid more attention to people,and gets preliminary application with its passive,long transmission distance,resistance to electromagnetic interference and so on.The correct extraction and events recognition of optical fiber sensing information is one of the key indicators.In this paper,an intrusion signal extraction and identification algorithm is proposed.Extract intrusion fiber optical signal by using signal envelope extraction based on Hilbert transformation,which is combined with erosion and expansion operation in mathematical morphology.Denoise signal by using wavelet shrinkage and translation-invariant algorithm.Extract features of intrusion signal based on " wavelet packet-energy " signal feature extraction strategy.Classify the intrusion signal by using ANN based on BP and RBF separately.The simulation results shows the effectiveness of this algorithm.
出处
《计算机技术与发展》
2014年第6期161-165,共5页
Computer Technology and Development
基金
国家自然科学基金资助项目(61272084)
江苏省高校自然科学研究重大项目(11KJA520002)
高等学校博士学科点专项科研基金资助课题(20113223110003
20093223120001)
关键词
分布式光纤入侵检测系统
小波除噪
平移时不变算法
人工神经网络
distributed fiber optical perimeter intrusion detection system
wavelet denoising
translation-invariant algorithm
ANN