Design And Implementation Of Embedded Remote Monitoring System Based On ARM9

Video surveillance system can transmit the field information to the monitoring clienttimely and rapidly, so it is widely used in buildings, office buildings and other places.Through analyzing the common monitoring systems of domestic and foreign, we can foundthat embedded monitoring platform is stable, high performance and low powerconsumption, it can realize image acquisition, encoding, decoding and networktransmission effectively, it will be the direction of video surveillance systems development.Therefore, it has important practical significance to carry out research and application forvideo surveillance technology based on embedded platform. In this paper, after the researchand analysis of embedded ARM9microprocessor S3C2410chip functions and Linuxoperating system, we designed a video surveillance system and made it realized. The workmainly includes the following:(1) Designed an embedded remote video surveillance system based on ARM9operating system, gave the overall system design and framework and than described thesystem work principle and processes.(2) Discussed S3C2410, memory SDRAM and design of interface circuit of Flash.Data acquisition, transmission and other functions are used in the monitoring system, so weanalyzed the S3C2410, chip card DM9000, chip RS232and USB interface circuit design.(3) Introduced the software development process of embedded system, described thecomposition, organization and configuration of the Linux kernel in detail, and then, weconfigured the kernel in order to make it support for the S3C2410board. In addition, wedescribed the methods of create the embedded system files, and created the necessary andwe required file systems using Busybox software.(4) In order to ensure the normal use of the camera, we added the USB camera driverin the Linux kernel successfully, and then realized the image data acquisition follow theVideo4Linux process.(5) On the base of JPEG compression algorithm analysis, it contains a large amount ofcomputation on DCT transformation, so we improved the algorithm. After analyzed thequantified data, the improved ideological is converts the2D DCT to two consecutive one-dimensional DCT, and then using the AAN fast algorithm. This method cansubstantially reduce the computation. Compared the compressed reconstructed pictures, wecan found that the improved method not only reduced the computation, but also does notaffect the quality of the pictures.(6) On the basis of analysis and comparison of TCP and UDP communicationprotocols, we used the UDP protocol socket for the normal communication between theserver-side (ARM9platform) and the client (PC), the client displays images after thereceived data decompressed.