| Detecting and tracking of moving objects are two key techniques of intelligent videosurveillance systems. Meanwhile video acquisition processing compressing, andtransmitting techniques are needed too.This thesis implements a complete intelligent video surveillance platform which usesgeneral-purpose embedded system as video capture and transmission terminal, and PCas video display and processing server. The real-time video data transmitting can beachieved by connecting terminal and server via the Ethernet.The overall design scheme is proposed by the thesis based on the functionalrequirements of video surveillance system. As to embedded system developmentprocess, S3C2440chip hardware platform, the USB interface, network interface andother resources are discussed. The construction of embedded Linux operation system,application development method and video capture technology on Linux system areresearched. The TCP/IP SOCKET programming on both Linux and Windows platformis analyzed. As to PC part, the features and applications of the visual processing libraryOpenCV are researched. The dynamic video detection methods of inter-frame differencemethod, based on optical flow technology method, background subtraction method arediscussed.The thesis implements USB camera video data acquisition, image data formatconversion and network interface software design. The video capture and video networktransmission on embedded Linux environment are archived. The sever bases on VC++and OpenCV software platform, using visual processing library and computervision-related functions, through inter-frame difference method implement a dynamicvideo surveillance system.For the future application expansion and increasing video surveillance node, thethesis divides the system into terminal and server implementations. This thesis has somereference value on embedded system applications and video object detection. |
PDF Full Text Request