基于ARM11的网络视频监控系统设计

发布时间：2018-05-25 04:16

本文选题：嵌入式系统 + linux　；参考：《浙江师范大学》2014年硕士论文

【摘要】：网络监控主要是通过网络接收远端的网络摄像头传输的视频流,通过解码、显示、存储等,实现存储、预览、报警等功能。现阶段,市场上的网络监控主要是在NVR(网络硬盘录像机)、IPC(网络摄像机)、超大容量硬盘(一般是3块以上3T硬盘)、液晶显示器组成的硬件框架上,以网络传输技术、GUI(图形用户界面)技术、视频编解码技术、嵌入式技术及存储技术为核心,通过有线网络传输数据来实现整个方案。整个方案的体积大,可移动性差,布线困难且价格昂贵；其后端NVR一般用高端的ARM A9双核架构设计,前端IPC采用ARM9架构设计。由于双核的A9芯片价格较高,一般用在对视频质量要求较高的场合,如政府、超市、学校等。而针对家庭等视频质量要求较低的场合,这类视频监控还较少。有一些用ARM9芯片做后端的低端产品,性能又很有限。现阶段用ARM11架构做视频监控的,国内就更少了。目前,ARM9架构的芯片与ARM11架构的芯片在价格上已经相差不大,然而ARM11架构的芯片比ARM9架构的芯片,在功能上要强很多。在此背景下,本文选取ARM11架构中的S3C6410芯片,以linux操作系统为平台,支持通过WIFI传输网络数据,实现基于ARM11的嵌入式网络视频监控系统。该系统与ARM9架构的视频监控系统相比较,其成本相差不大、体积较小、可方便移动。本文分析了选择S3C6410作为视频监控系统主控芯片的优势,并根据视频监控的功能需求,选取了飞凌公司的0K6410开发板为硬件平台。由于嵌入式网络视频监控系统功能模块较多,整个系统比较庞大,本文简洁的介绍了其中一些非关键功能模块的实现流程,详细介绍了研究的关键模块。本系统建立了嵌入式交叉编译环境及裁剪linux内核,移植了Tslib、FreeType、FFmpeg三个开源软件库,开发了两个linux设备驱动程序：TFT触摸屏触摸输入功能设备驱动程序,和实现显示功能的帧缓冲设备驱动程序。现阶段,针对大部分嵌入式视频监控系统都是采用16个bit位显示一个像素的颜色值,该系统开发了一个像素采用32个bit位的显示设备驱动程序,这样能够显示的颜色值会大量增加,图像的失真也会相应减少。另外,对于图形用户界面技术,市面上已经有QT、miniGUI等功能完备的GUI技术,虽然这些GUI也有免费的版本(学习或交流),但用在商业上,都要收取授权费用的。基于以上这点,本系统开发了一个简单的GUI,该GUI利用帧缓冲设备控制图像在液晶屏上的显示,使用Tslib库获得触摸屏的输入信号,以及使用FreeType库获取显示文字的位图信息,实现了用户和机器的交互。另外,本系统还设计了一套信号与槽机制,利用这种机制能够有效处理GUI页面与页面、页面与控件之间复杂的信息传递过程。还值得一提的是,本系统开发了采集摄像头视频数据的应用程序,该应用程序在linux下的V4L2(video for linux two)摄像头驱动程序框架下运行,实现了视频数据实时采集的效果。最后,通过FFmpeg库对视频数据进行编码,降低了传输视频数据的网络带宽要求,同时又减少了录像时占用的大量I/O资源。系统用简单的往存储设备写文件的方式实现抓图、录像功能。在此基础上结合网络视频数据的传输实现整个嵌入式网络视频监控系统。
[Abstract]:Network monitoring is mainly through the network receiving video streaming of the remote network camera, through decoding, display, storage and so on, to realize storage, preview, alarm and other functions. At present, the network monitoring in the market is mainly in NVR (network hard disk recorder), IPC (network camera), super large capacity hard disk (generally more than 3 3T hard disk), liquid crystal display On the hardware frame of the indicator, the whole scheme is realized by network transmission technology, GUI (graphical user interface) technology, video codec technology, embedded technology and storage technology. The whole scheme is large, the mobility is poor, the wiring is difficult and the price is expensive; the back end NVR generally uses high-end A RM A9 dual core architecture design, front-end IPC using ARM9 architecture design. Because the dual core A9 chip is high in price, generally used in the video quality requirements, such as the government, supermarket, school and so on. But for the home and other video quality requirements of low level, this kind of video surveillance is still less. There are some low-end products with ARM9 chips as the back end. The current stage with the ARM11 architecture for video surveillance is less domestic. At present, the chip of ARM9 architecture and ARM11 architecture have little difference in price, but the chip of the ARM11 architecture is much stronger than the ARM9 architecture. In this context, this paper selects the S3C6410 chip in the ARM11 architecture to Linux. The operating system is the platform to support the transmission of network data through WIFI to realize the embedded network video monitoring system based on ARM11. Compared with the video monitoring system of the ARM9 architecture, the system has little difference in cost, small volume and convenient to move.
This paper analyzes the advantages of selecting S3C6410 as the main control chip of the video monitoring system. According to the functional requirements of video surveillance, the 0K6410 development board of the flying company is selected as the hardware platform. As the function modules of the embedded network video monitoring system are more and the whole system is huge, some of the non key functions of the system are briefly introduced in this paper. The key module of the energy module is introduced in detail. The system establishes the embedded cross compiler environment and the Linux kernel, and transplants the three open source software libraries of Tslib, FreeType and FFmpeg, and develops two Linux device drivers, TFT touch screen touch input device driver, and the frame to realize the display function. At this stage, most embedded video surveillance systems use 16 bit bits to display the color values of one pixel. The system develops a display device driver with 32 bit bits in pixels, so that the color values can be increased and the distortion of the image will be reduced. Graphical user interface (GUI) technology has QT, miniGUI and other complete GUI technologies. Although these GUI have free version (learning or communication), they all have to charge authorization costs in business. Based on this, this system develops a simple GUI, which uses a frame buffer to control the display of the image on the LCD screen. It uses the Tslib library to obtain the input signal of the touch screen, and uses the FreeType library to obtain the bitmap information of the display text, and realizes the interaction between the user and the machine. In addition, this system also designs a set of signal and slot mechanism, which can effectively handle the complex information transfer process between GUI page and page, page and control. It is also worth mentioning that the system has developed an application program to collect video data from the camera, which runs under the framework of V4L2 (video for Linux two) camera driver under Linux, and realizes the effect of real-time video data acquisition. Finally, the video data is encoded by the FFmpeg library, and the video data is transmitted. Network bandwidth requirements, while reducing a large number of I/O resources occupied by video recording, the system uses a simple way to write files to the storage device to realize the capture and video function. On this basis, the whole embedded network video monitoring system is realized with the transmission of network video data.
【学位授予单位】：浙江师范大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN948.6

【参考文献】