摘要
介绍了自主研发的一种针对计算机网络电磁泄漏信号的自动化实时监测系统,包括电磁信号接收装置和电磁信号监测平台两部分,能够实现对目标设备电力线上的泄漏信号的实时监测与信号的模板采集功能。并以此为平台,借助深度学习与图像识别技术对电磁泄漏信号进行了分类识别,验证了卷积神经网络应用于电磁泄漏信号识别的有效性,为后续该领域的相关研究提供了思路。经过测试与分析,发现识别结果会受到样本训练集种类数、接收装置探头的位置以及样本训练集间的相似度影响,这对提升电磁泄漏信号的分类识别能力提出了更高的要求。计算机网络电磁泄漏信号的监测与智能识别的研究对电磁泄漏源的精准定位以及电磁信息泄漏威胁程度的评估有着重要意义,也为电磁泄漏的预警与防护提供了重要依据。
a self-developed automatic real-time monitoring system for electromagnetic leakage signals in computer network is introduced,including electromagnetic signal receiving device and electromagnetic signal monitoring platform.It can realize real-time monitoring of the leakage signal on the power line of the target device and template acquisition of the signal.Using this as a platform,the electromagnetic leakage signal is classified and identified by deep learning and image recognition technology,which verifies the effectiveness of convolutional neural network in electromagnetic leakage signal recognition,and provides ideas for subsequent research in this field.After testing and analysis,it is found that the recognition result will be affected by the number of sample training sets,the position of the receiving device probe and the similarity between the sample training sets,which puts forward higher requirements for improving the classification and recognition ability of electromagnetic leakage signals.The monitoring and intelligent identification of electromagnetic leakage signals in computer networks is of great significance for the accurate positioning of electromagnetic leakage sources and the assessment of the threat level of electromagnetic information leakage,and also provides an important basis for the early warning and protection of electromagnetic leakage.
作者
王梦寒
寇云峰
刘文斌
兰宇
程磊
李雨锴
WANG Meng-han;KOU Yun-feng;LIU Wen-bin;LAN Yu;CHENG Lei;LI Yu-kai(Chengdu Xinxinshenfeng Electronic Technology Co.,Ltd.,Chengdu Sichuan 611731,China;China Cyber Security Co.,Ltd.,Chengdu Sichuan 610041,China)
出处
《通信技术》
2019年第7期1755-1760,共6页
Communications Technology
关键词
电磁泄漏
实时监测
卷积神经网络
智能识别
electromagnetic leakage
real-time monitoring
convolutional neural network
intelligent identification