Research And Implementation Of Web Page Tamper Detection System Based On Video

With the rapid development of network information technology,the issue of website security has received more and more great attention from the state and government.In order to improve the ability of relevant departments to respond to website security incidents,offensive and defensive rehearsal activities have come into being.In the process of offensive and defensive drills,the defending side needs to arrange professional personnel to be on duty to check webpage tampering at all times,which is very labor-intensive.In addition,the current web page tampering detection system uses web crawlers to obtain web pages,which leads to the system being exposed to the network and also becomes an attack target,which in turn affects the accuracy of subsequent tampering detection results.Therefore,a video-based web page tampering detection system is proposed.In response to the problem of poor security in web crawler technology for acquiring web pages,a video-based positioning method for web page acquisition is proposed.Firstly,the webpage video signal is reduced to webpage image,and then ECA-DOC-Res Net50 feature extraction network,adaptive interval grouping method and IOM post-processing screening strategy are designed to solve the problem of low accuracy of webpage text detection by CTPN algorithm.The text detection results are fed into the CRNN recognition network to extract the text content.Finally,based on the detected text location information,a web page traversal strategy is developed and a simulated mouse is used to achieve automatic web page acquisition,which ensures the security of the web page tampering detection system.In response to the problem that the web page tampering detection algorithm has a single detection type and poor detection effect,a web page tampering detection algorithm based on multimodal fusion is proposed.Firstly,coordinate attention mechanism is added to Mobile Net V2 network to improve the attention to image features;then text features are extracted using ALBERT model,and finally,transformer structure is introduced to construct multimodal fusion model for web page tampering detection.The effectiveness of this model is verified by comparing with the single-modal model and the traditional fusion model.Finally,the embedded system hardware platform is built,and the video-based webpage tamper detection system is developed,which realizes the automatic webpage acquisition function and webpage tamper detection function.Using this system reduces the cost of website security detection and improves the security and accuracy of web page tampering detection system.