Research And Implmentation Of IoT Device Recognition Based On Deep Learning

With the continuous maturity of technologies such as 5G,big data,cloud computing,and edge computing,IoT is gradually becoming the core driving technology of a new round of technological revolution,and the number of IoT devices connected to the Internet is also growing rapidly.However,the continuous access of large-scale IoT devices also brings a series of security risks.Once the IoT devices have loopholes,it may cause serious security problems.Therefore,it is of great significance to effectively identify IoT devices in the network environment.Existing IoT device identification methods have problems such as large recognition granularity and incomplete information extraction,but the rapid development of deep learning provides new ideas for effectively extracting device information.Therefore,this paper applies deep learning technology to IoT device identification,and designs and implements a solution for IoT device identification and vulnerability discovery.Specifically include:1.The thesis proposes a method for IoT device identification based on convolutional neural network.This method makes full use of the nontext features of the Internet of Things device messages.Firstly,extract the icon of the background landing page of the Internet of Things device,and then extract the image feature vector from the icon based on the method of convolutional neural network,and use the feature vector to build the fingerprint library of the Internet of Things icon.Finally,realize the identification of connected device brands by means of image retrieval.2.The thesis proposes a BERT-BiLSTM-CRF-based information extraction method for IoT devices.This method uses BERT in the process of extracting IoT device information,and performs named entity recognition on the IoT device information in the search results,so as to obtain device information of other attributes according to the model of the IoT device.3.Based on the IoT device identification method mentioned above,combined with existing research,an IoT device identification and vulnerability discovery system based on multi-feature fusion is designed and implemented.The system can comprehensively identify the type,brand,and model information of the equipment from various characteristic information of the IoT equipment,and combine the existing public vulnerability database of the IoT equipment to realize timely alarm for security issues related to the IoT equipment.To further ensure the security and stability of the network environment.