The Design And Implementation Of Intelligent Network Video Camera Based On S3C2440

Video surveillance is a very important part of Security System. From the closed-circuit television to the network video surveillance, video surveillance technology have changed a lot. In the analog video surveillance phase, the camera connect to monitor with proprietary cable,every camera need a monitor,and Because the cable length is limited to a smaller range surveillance. Subsequently, the computer processing technology energizing the video surveillance , video surveillance has gradually changed from analog to digital surveillance monitoring, greatly enhanced functionality of video surveillance, expanding the video surveillance applications. Emerged in recent years, network video surveillance aircraft out of dependence on the host computer, using embedded technology, network transmission of video, each network video monitoring machines as an independent individual accepte customer's request.Network video surveillance have the advantage of easy to deploy and can monitor a wide range. The main work of this paper is realization of a network video surveillance camera based on ARM9 S3C2440 chip. Most of the current network video surveillance monitor on the market only have a single function, still require specialized staff to view live video or video recording, and hard to avoid oversight. Network video camera by adding in moving object detection and tracking algorithm can carry out a alarm through the client-side to alert monitoring people may be a potential threat when the camera detect a moving object.Then the monitor people watch the monitor to eliminate the threat. The video monitor also have Web-based video services, the client can watch video through the browser. The system also have intelligent analysis client with intelligent processing algorithms can deal with the detected moving object. System is divided into two parts, hardware and software. The hardware part based on ARM9 platform through the decision of the ADI DSP and ARM9 platform, and the embedded operating system use embedded Linux operating system which have the support of open-source community. Embedded platform using SDRAM as the memory, FLASH as external storage devices, FALSH including the NOR FLASH 2M and 64M of NAND FLASH. NOR FLASH with the advantages of a lower error rate and running the program directly, usually to store the key code, this article store BootLoader program using NOR FLASH, store the Linux kernel and applications using NAND FLASH.Thus system can wirte NAND FLASH throngh Bootloader when the camera need upgrade. BootLoader using open-source U-boot, compared to other BootLoader, U-boot support better function to the developer. Because U-boot must be used in the hardware platform desicribed in this article, we need to modify the configuration file. In order to simplify the migration process, using SMDK2410 development board as the sample and then modify the configuration file. The CPU in SMDK2410 is the S3C2410,the main difference between S3C2410 and S3C2440 is frequency mode and the control type of NAND FLASH. We only need to modify this part of the code and configure the correct SDRAM, FLASH and other hardware models then U-boot can running on the camera hardware platform. Linux kernel is the basis of system operation, to use the make menuconfig command to configure the kernel so that it can adapt to S3C2440 platform configuration, and modify the frequency of the hardware parameters, and then packaged into a U-boot the kernel image file format, then write to the FLASH. In addition, we need to partition the FLASH,the pre-2M is used to store core, the remaining space used to store the file system and application executable file. Study the building process and key technologies of Intelligent control system,put forward a modular method of achieving monitor intelligent through embedded platform and general-purpose computer platform. In the camera, video capture device include external USB camera and CMOS camera integrated in the main board. USB camera have the advantages of easy to use, easy procurement, but also easy to upgrade. CMOS camera module using OmniVision's OV9650 chip, this article prepared for the chip standard driver to follow Linux V4L interface, V4L is the Linux kernel interface to support video processing, provide a bridge between video processing applications and hardware for the interaction. Video capture module, using V4L provides a variety of data structures and access interface capture video data. Taking into account the embedded computing platform of limited resources and the accuracy selecting from a wide range of moving target detection algorithm the method of the dynamic background update and background subtraction to detect the moving object. Similarly, the target tracking algorithm using color histogram method combined with Euclidean distance. In addition, the open source Web server Boa transplanted to the camera, using CGI programs and JavaApplet technology support for the browser, so users can view the camera video through the browser, real-time surveillance video. In addition to the browser, but also through a dedicated client, a dedicated customer has a more powerful function, can provide customers with intelligent alarm and further analysis. In order to reduce the amount of transmission video data, the video data need to be encoded, but because of the limitations of the embedded platform, if using overly complex the coding algorithm, it is difficult to maintain the smooth flow of video and ultimately choose the MJPEG encoding algorithm sd the video compression algorithm. The client receives a moving object from the server,and analysis the object,intelligent analysis includes two parts classification module and content-based retrieval module. Classification of moving objects based on five features,vehicle and mankind can be classified. Content retrieval module put the object image into the image database to match, depending on the application, create a different image database. System is expected applications include residential security, prison security systems, alarm against moving targets. Identification of regional disasters, mainly for content-based retrieval system. In the future camera can also increase the SIM module, by sending a text message or call reminder of the potential threat to the user; Independent power supply module for the camera, so the camera can be deployed more convenient. GPS module for the camera, enabling the system at any time to determine the camera position, the camera appears to prevent migration. To increase the infrared camera module, so the camera can work at night or low light situations.