Research On Traffic Identification Technology Of Video Capture Devices In Encrypted Wi-Fi Environment

Nowadays,Wi-Fi is almost everywhere in people's life and work,and encryption technology based on 802.11 protocol provides a guarantee for the security and reliability of communication transmission.However,in reality,there exists the phenomenon that some malicious users use hidden video capture devices to transmit videos to the outside in real time by encrypted Wi-Fi.These devices transmit others' private information or confidential information of enterprise organizations "in invisible",and because of the encryption of communication,the transmitted information is not easy to be detected by normal users and network administrators,which leads to the disclosure of important information.In order to solve this problem,this paper proposes a method to identify whether the traffic statistical characteristics of each communication device accord with the traffic statistical characteristics of video capture activities in the encrypted Wi-Fi network,which can judge whether these devices are video capture devices.In this study,under the condition of no decryption of encryption,the traffic statistical characteristics of video capture activities and other communication activities are analyzed in the encrypted Wi-Fi environment,and Support Vector Machine(SVM)is used to model learning to obtain classifiers for identifying the traffic of video capture activities based on the difference of traffic statistics characteristics.Firstly,the traffic characteristics of video capture activities and other communication activities in encrypted Wi-Fi environment are analyzed statistically,and it is found that there are obvious differences between the traffic of video capture activity and other communication activities in rate,frame length,frame distribution of uplink and downlink.The feature selection algorithm combined with CFS is used to select the best feature subset for traffic identification from the candidate feature set.Then the traffic identification optimization method of video capture devices based on SVM is designed and implemented.Considering that there are abnormal samples in the training samples of this study,a training samples selection algorithm is applied to the identification method of this study,which improves the rate and effect of the classifier,and it is better than the classifiers based on other machine learning methods.Finally,in order to judge whether the communication device in the encrypted Wi-Fi environment is avideo capture device,a traffic identification system for video capture equipment is designed and implemented based on the traffic identification classifier.It is verified that the system can detect the video capture devices quickly and accurately in the real Wi-Fi environment.The accuracy of the traffic identification classifier is as high as 97.04%,and the false positive rate is only about 3.43%.Based on the classifier,the time of which the video capture devices are detected by the system is only about 18 seconds.