基于深度学习网络的光通信系统非法入侵行为识别研究

Research on identification of illegal intrusion behavior in optical communication system based on deep learning network

为了检测多种光通信系统非法入侵行为,保障光通信系统运行安全,提出了基于深度学习网络的光通信系统非法入侵行为识别方法。采用光栅传感技术检测光栅传感器反射波长偏移量,感知光通信系统非法入侵行为频率特征信号。利用小波包分解方法将频率特征信号分解成多个频带后,提取各频带小波包能量,将其作为卷积神经网络输入。经小波包能量特征提取、处理、融合操作后,通过Softmax分类器完成光通信系统非法入侵行为数据特征的分类,实现光通信系统非法入侵行为识别。实验证明:该方法可迅速挖掘出光通信系统中非法入侵行为的时域、频域特征信号。所提取小波包能量可准确反映光通信系统中非法入侵行为特点。该方法可实现多种光通信系统非法入侵行为精准识别,助力管理人员针对入侵行为做出对应防御措施。

This paper studies the identification method of illegal intrusion behavior of optical communication system based on deep learning network,detects various illegal intrusion behaviors of optical communication system,and ensures the operation safety of optical communication system.The grating sensing technology is used to sense the frequency characteristic signal of illegal intrusion behavior of optical communication system by detecting the reflected wavelength offset of grating sensor.The wavelet packet decomposition method is used to decompose the frequency characteristic signal into multiple frequency bands,extract the wavelet packet energy of each frequency band,and input it as a convolution neural network.After the wavelet packet energy feature extraction,processing,and fusion operations,The Softmax classifier is used to complete the classification of illegal intrusion data characteristics of the optical communication system and realize the identification of illegal intrusion behavior of the optical communication system.The experimental results show that this method can quickly mine the time-domain and frequency-domain characteristic signals of illegal intrusion in optical communication systems,and the extracted wavelet packet energy can accurately reflect the characteristics of illegal intrusion in optical communication systems;This method can realize accurate identification of illegal intrusion behaviors of various optical communication systems,and help managers to take corresponding defense measures against intrusion behaviors.