大规模在线监控设备的隐私及安全检测研究

发布时间：2018-01-29 12:06

本文关键词： 网络测量物联网设备视频监控设备信息物理系统　出处：《北京交通大学》2017年硕士论文　论文类型：学位论文

【摘要】：物联网是我国战略性新兴产业的重要组成部分,引领了继计算机、互联网和移动通信之后的新一轮信息技术革命,是未来科技竞争的制高点和产业升级的重要驱动力,是加速推进工业化、信息化融合的催化剂。目前,越来越多的物联网设备出现在网络空间内,包括网络摄像头,工业控制设备,智能家电等等设备。随着物联网行业的发展,网络空间内的物理设备逐渐将自身暴露在公共区域,这既推动了社会的发展,也带来了安全与隐私问题。网络空间内可被直接观测到的视频监控设备日益增加,这直接导致大量用户的隐私安全受到威胁。在本篇论文中,提出了一种可以实时监控大规模视频监控设备安全及其隐私泄露情况的方法。设备搜索方面,本系统采用单包探测及无状态扫描架构对全网内近四十亿个网络地址进行检测,利用异步探测架构可以实现实时定位大规模范围内的视频监控设备。针对待测目标集合,根据目标设备的视频传输协议,利用HTTP及RTSP协议所反馈的标语信息对服务端设备身份进行识别。结合自然语言处理模块对经过预处理的标语信息进行分类及标定,提取设备的厂商及品牌信息以达到识别目标设备身份的目的。在获取到当前网络空间内高危视频监控设备的网络地址后,系统会调用隐私泄露检测模块对待测目标进行安全检测。本系统以目标设备用户的隐私权益为出发点,采用计算机视觉识别技术对所泄露的场景信息进行提取,只存储设备所监控区域的场景标签,不存储任何来自设备的音视频流,最大程度上的保护用户的隐私安全。本研究提出了一种视频监控设备的实时搜索与识别方法,同时根据该方法设计并实现了大规模网络空间监控设备的隐私泄露检测系统原型。本系统对当前网络环境内的监控设备品牌及型号进行了综合性探测,勾勒了设备品牌及类别的全景图,并对当前市场内普及的视频监控设备进行了总体性的脆弱性分析及隐私泄露程度检测。实验方面,系统根据国家、城市及网络服务运营商等筛查标准,对全网空间内监控设备进行了综合性多层次的扫描,确定并定位了安全漏洞的分布及严重情况。该工作弥补了当前对于实时监控视频监控设备方面工作不完善的空白,实验也为物联网领域内相关工作的进一步研究提供了数据样本。
[Abstract]:Internet of things is an important part of China's strategic emerging industries, leading a new information technology revolution after computers, the Internet and mobile communications. It is an important driving force of science and technology competition and industrial upgrading in the future. It is the catalyst of accelerating industrialization and information integration. At present, more and more Internet of things devices appear in the cyberspace. With the development of the Internet of things, the physical devices in the cyberspace gradually expose themselves to the public area, which has promoted the development of the society. Also brings security and privacy issues. In cyberspace can be directly observed video surveillance equipment increasing, which directly leads to a large number of users' privacy security is threatened. In this paper. This paper presents a method to monitor the security and privacy of large-scale video surveillance equipment in real time. The system uses single-packet detection and stateless scanning architecture to detect nearly 4 billion network addresses in the whole network. The asynchronous detection architecture can be used to locate the video surveillance equipment in a large scale in real time. According to the target set, the video transmission protocol of the target device can be used. Using the slogan information feedback from HTTP and RTSP protocol to identify the identity of the server device, and combining with the natural language processing module to classify and calibrate the pre-processed slogan information. Extract the equipment manufacturer and brand information to achieve the purpose of identifying the target device. After obtaining the network address of the high risk video surveillance equipment in the current cyberspace. The system will call the privacy leak detection module to test the target for security detection. This system takes the privacy rights and interests of the target equipment user as the starting point and uses computer vision recognition technology to extract the leaked scene information. Only scene labels for the area monitored by the device are stored, and no audio and video streams from the device are stored. This paper presents a real-time search and recognition method for video surveillance equipment. At the same time, the prototype of the privacy leak detection system of the large-scale cyberspace monitoring equipment is designed and implemented according to this method. This system comprehensively detects the brand and model of the monitoring equipment in the current network environment. It outlines the panoramic picture of the brand and category of the equipment, and carries on the overall vulnerability analysis and the privacy leak degree detection to the video surveillance equipment which is popularized in the current market. Experimental aspect, the system according to the country. City and network service operators and other screening standards, the whole network of monitoring equipment in the space of a comprehensive multi-level scanning. Determine and locate the distribution of security vulnerabilities and serious situation. This work makes up for the current work of real-time monitoring and video surveillance equipment is not perfect blank. The experiment also provides data samples for further research in the field of Internet of things.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP309

【相似文献】