| In recent years, our country proposed the plan of "eye in the sky project". Namely by means of cloud computing technology, all of the country's video surveillance systems will be centralized management and monitoring. Combined with image processing and pattern recognition technology, we can achieve functions of intrusion detection, locating people, regional human traffic statistics, identifying traffic violations, fatigue driving detection, crime investigation and so on.This paper designs and realizes a cloud video surveillance system based on H.264. This system uses Samsung Exynos4412 embedded chips and embedded Linux operating system as the hardware and software development platform of the system. The video capture is designed based on Linux V4L2. The video encoding module is based on the current mainstream H.264 video codec standard of the surveillance system which codes the original video data reasonably. The video transmission module uses the Real time streaming protocol(RTSP) and Real-time transport protocol(RTP). RTP protocol is responsible for streaming the coded data to the cloud server and RTSP protocol is responsible for exchanging the interaction data between the server and front-end equipment during video data transmission. Compared with traditional video surveillance systems, this system does not require the deployment of local video processing devices and processes real-time video streaming in the cloud servers by the network transmission of video data. Cloud servers can decode, display and storage the video data and users can also access and control through PC or mobile terminals.According to the test result, real-time video-delay is controlled in 500 ms or less. Besides, there is little video jitter and data loss. Therefore this cloud video surveillance system is relatively satisfactory and achieves the expected performance; basically meet the average performance of cloud video surveillance systems. This video surveillance system has broad market prospect. |
PDF Full Text Request