基于SIP协议和视频摘要技术的考场监控系统研究

发布时间：2018-03-08 03:25

本文选题：SIP　切入点：考场监控系统　出处：《山东师范大学》2017年硕士论文　论文类型：学位论文

【摘要】：考试是选拔人才的通用途径,种类繁多。为了确保考试的公正性与透明性,考场监控系统应运而生。传统的考场监控系统效率极低,为了实现考场监控视频的快速浏览、及时获取有用的信息、锁定可疑的作弊对象,对传统的视频进行摘要变得尤其重要。SIP协议作为应用层中重要的协议,能够实现多媒体数据之间的相互建立或者终止,具有良好兼容性以及扩展性。因此,将SIP协议与视频摘要技术应用于考场监控系统中有着巨大的研究意义。本文主要工作以及创新点如下:(1)将SIP协议、RTP/RTCP协议进行了分析,然后通过SIP协议使UA和代理服务器之间进行相应的对话,并借助Wireshark和Free Switch对SIP协议会话方式和流程进行了分析。(2)通过SIP摄像头获得监控视频,将该系统的技术体系、实现方式及主要功能进行介绍,使客户端能够通过SIP协议与服务器端进行互联互通,以此完成考场监控系统中各个部分之间的高效连接。并对视频摘要的相关概念与技术进行论述。(3)针对图像中所出现的噪声问题进行去噪处理,并对进行运动目标检测的背景相减法、帧差法以及光流法进行介绍。本文中,给出一种改进的高斯滤波的方法,对抑制噪声、平滑图像以及使图像边缘更加清晰具有很大的作用。并针对背景差分法中“漏检”情况的发生,提出了一种改进的混合高斯模型的方法,该方法能够增加检测的正确性。(4)在运动目标跟踪算法Meanshift和Camshift的基础之上,将RGB颜色空间与HSI颜色空间进行转换。HSI色彩空间相对于RGB色彩空间而言对颜色的感知更加直观,解决了RGB颜色空间易受周围环境影响的缺陷。(5)设计了一种视频摘要系统算法,并利用C++与Open CV开源库结合其他相关库对考场中活动幅度较大的考生进行跟踪检测。
[Abstract]:In order to ensure the fairness and transparency of the examination, the examination room monitoring system emerges as the times require. The traditional examination room monitoring system is extremely inefficient, in order to realize the fast browsing of the examination room surveillance video, the examination is a common way to select talents. Getting useful information timely, locking down suspected cheating objects and abstracting the traditional video becomes particularly important. Sip protocol, as an important protocol in the application layer, can realize the mutual establishment or termination of multimedia data. It has good compatibility and expansibility. Therefore, it is of great significance to apply SIP protocol and video summary technology to the test room monitoring system. The main work and innovation of this paper are as follows: 1) the SIP protocol is analyzed. Then, through the SIP protocol, the UA and the proxy server can talk to each other, and with the help of Wireshark and Free Switch, the conversation mode and flow of SIP protocol are analyzed. (2) the surveillance video is obtained through the SIP camera, and the technical framework of the system is put forward. The implementation mode and main functions are introduced to enable the client to interconnect with the server through the SIP protocol. In order to complete the test room monitoring system between the various parts of the efficient connection. And video abstract related concepts and techniques are discussed. The background subtraction method, frame difference method and optical flow method for moving target detection are introduced. In this paper, an improved Gao Si filtering method is presented to suppress noise. Smoothing the image and making the image edge clearer has great effect. In view of the occurrence of "miss detection" in the background difference method, an improved mixed Gao Si model is proposed. This method can increase the correctness of detection. (4) based on the moving target tracking algorithms Meanshift and Camshift, the color space of RGB and HSI color space are transformed into. HSI color space, which is more intuitive to the color perception than the RGB color space. This paper solves the defect of RGB color space which is easily affected by the surrounding environment. It designs a video abstract system algorithm, and uses C and Open CV open source library combined with other related libraries to track and detect the examinees with large activity range in the examination room.
【学位授予单位】：山东师范大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP391.41

【参考文献】