| With the development of the society and the rising of the living standard and the industrial automation, the unexpected events between people and public infrastructure have become increasingly frequent. In addition, these incidents often have the characteristics of no symptoms and difficult to predict. In such a situation, video surveillance is becoming more and more important.FPGA-based processing technology has the advantages of stable performance, short development cycle, configuration flexibility and so on. Therefore, FPGA is used in this video surveillance system as a core processor.JPEG image encoding and decoding algorithm is the first international coding standards for still image and grayscale images.JPEG applies not only to the still image compression encoding, but also be used as a video image compression standard. Firstly, the research of the FPGA-based video image processing technology is studied, on this basis, a system of hardware and software platforms are designed. The hardware part includes the video capture module, the video coding module, the network transmission module, the network receiving module, the video decoder module, and the video display module. Software part includes JPEG image encoding and decoding algorithms and the LWIP network transport protocols. Since the system is relatively complex, the requirements of the FPGA chip speed and the internal logic resource are quite high, so the XILINX’s VIRTEX-Ⅱ PRO series XC2VP50 FPGA chip is used.The video surveillance system is divided into input and output terminals of two parts. At the input, the video image directly becomes from the camera, and the image data is compressed in the JPEG format by the video encoder module. At the output part, the compressed image data is obtained by the video decoding module and finally displayed on the video display terminal. For the real-time data transmission between input terminal and an output terminal, the Lwip network protocol stack embedded on the Xilinx FPGA is used, and the UDP transport form is used to finally get the original low-latency video data. |
