The Design And Implementation Of Software Platform Of Embeded Video Analysis System

Video surveillance is a system which has relatively strong ability to prevent the occurrence of crimes. It can provide the most intuitive and accurate information, its application in society is becoming increasingly widespread. Video surveillance has experienced analog video surveillance system, digital video surveillance system two stages, and is now gradually stepping to intelligent video surveillance system (IVS). Intelligent video surveillance systems rely on image processing techniques to detect anomalies automatically in the picture instead of simply relying on the human eye to observe. This monitoring system not only reduces the work pressure of staff, but also improves the effectiveness of monitoring and early warning.Video analysis system is the intelligent core of intelligent monitoring system, responsible for completing the detection and recognition of abnormalities in video sequence anomaly. Video analysis systems are generally based on specific embedded hardware platform, using the algorithms related to video processing to achieve the smart detection purpose.The object of this article is to complete the design of platform software, which is to transplant Linux operating system to the hardware platform and to complete systems booting. This paper mainly completed the following work: built a host-side development environment; transplanted UBL, U-Boot to the self-made hardware platform, realizing Linux kernel booting; configured kernel based on resources on board and built root file system using busybox. Established a friendly system environment for the following algorithm development; transplanted and developed peripheral device driver, making the Linux support all peripherals on hardware platform. Finally, wrote the kernel and file system to the NAND flash. A complete development platform for embedded video analysis system was constructed. Algorithm modules can be developed and debugged on the platform and built into Linux kernel, thus a suitable embedded video analysis system.The embedded video analysis system this thesis participate in has the advantage of flexibility, Rich interface, algorithm modules which is developed, updated and uninstalled independently. This system is suitable for such occasion that has relatively many and scattered monitoring points and raise higher requirements for video quality, need to remote monitoring.