基于嵌入式Linux的网络视频监控系统的研究与设计

发布时间：2018-12-21 20:25

【摘要】：随着大数据和云计算时代的到来，互联网的越发成熟并广泛的普及于人们生活的方方面面。视频监控作为对未来平安城市、智慧城市的重要支撑手段，也越发显得尤为重要。本课题研究并设计一种基于嵌入式Linux的网络视频监控系统，整个系统采用B/S架构，即：以三星公司的S3C2440作为视频监控的核心处理器，通用计算机作为监控显示和控制终端，利用客户端浏览器直接访问服务器，最终在PC机上显示监控画面。首先在开发板搭建嵌入式Linux系统平台，，依次移植U-Boot引导程序、Linux内核的移植和定制、创建根文件系统，并在此基础上进行网络视频监控系统的设计。网络视频监控系统由五大模块组成，分别是视频采集模块、视频压缩模块、视频传输模块、Web服务器模块以及客户端模块。视频采集模块利用Linux提供的V4L2应用程序接口进行视频数据的采集；然后经过H.264视频压缩模块进行视频数据的压缩；接着利用流媒体传输协议传输视频数据；客户端可以请求Web服务器，Web服务器接收请求后，将视频数据反馈给客户端。网络监控系统设计完成之后，又对三种移动目标跟踪算法做了相关研究。MeanShift算法的跟踪窗口不能根据跟踪目标大小自适应的调整；CamShift算法虽然能自动调节跟踪窗口大小来准确跟踪目标，但是如果目标颜色和背景颜色相近或是移动目标被完全遮挡，将导致丢跟踪失败；基于Kalman滤波改进的CamShift目标跟踪算法，通过时间更新和测量更新预测下一帧移动目标可能的位置，很好的解决颜色相似或是移动目标被完全遮挡的情况，能够在较复杂的环境下准确跟踪目标。
[Abstract]:With the arrival of big data and cloud computing era, the Internet has become more mature and widely used in all aspects of people's lives. Video surveillance as an important means of support for the future safe city, intelligent city, is also becoming more and more important. This paper studies and designs a network video surveillance system based on embedded Linux. The whole system adopts B / S architecture, that is, Samsung S3C2440 is used as the core processor of video surveillance, and the general computer is used as the monitor display and control terminal. The client browser is used to access the server directly, and finally the monitor screen is displayed on the PC machine. Firstly, the embedded Linux system platform is built in the development board, then the U-Boot boot program is transplanted, the Linux kernel is transplanted and customized, the root file system is created, and the design of the network video monitoring system is carried out on this basis. The network video monitoring system is composed of five modules, namely, video capture module, video compression module, video transmission module, Web server module and client module. The video acquisition module uses the V4L2 application program interface provided by Linux to collect video data, then compressing video data through H.264 video compression module, and then transmits video data by streaming media transmission protocol. The client can request the Web server, and the Web server can feedback the video data to the client after receiving the request. After the design of the network monitoring system has been completed, three moving target tracking algorithms have been studied. The tracking window of the MeanShift algorithm can not be adjusted adaptively according to the size of the tracking target. Although the CamShift algorithm can automatically adjust the size of tracking window to track the target accurately, if the color of the target and the background color are similar or the moving object is completely occluded, it will lead to the loss of tracking failure. The improved CamShift target tracking algorithm based on Kalman filter can predict the possible position of moving target in the next frame by time update and measurement update, which can solve the problem of similar color or the moving target being completely occluded. Able to track targets accurately in more complex environments.
【学位授予单位】：中国矿业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN948.6

【共引文献】