A Study Of Moving Vehicle Detection And Tracking System Based On DM6446

With the rapid development of modern road traffic and continuous improvement to traffic monitoring requirements, the video surveillance system in intelligent transportation and other fields fields plays an increasingly important role, and develop in a real-time, miniaturization direction. Monitoring system combines video and computer uses a lot of machine vision and image processing algorithms to monitor the traffic and transport situation, get traffic information and provide useful data to slove many problems due to the rapid development of transport caused. Goal of this study is to detect and track moving vehicles in a complex traffic environment using Open CV vision library for program development, and porting to DM6446 embedded platforms used in practical environments.In practice, the traffic environment requires synchronization of data processing and data acquisition, higher hardware processing speed. Taking into account the above factors, this project selected TI technology ’s Davinci DM6446 development platform based on dual-core architecture. It not only meets the requirements of real-time development, and also is a dedicated product for vido applications, it reflects the advantages of the ARM and DSP processors on the control performance and video processing speed.Therefore TMS320DM6446 can play a unique performance advantages in this project. The main research work includes the following aspects:(1) Depth understanding of the DM6446 software, hardware architecture, studying and testing the system development framework, building Arm_v5t cross-compiler environment, NFS, TFTP, minicom and other system services and applications in Linux operation system, compiling kernel file system and development board bootloader running on ARM, also building a detailed software and environmental for development board debugging(2) Studying TI algorithm specifcation includes Codec Engine, Codec Server, XDM and XDAIS, etc., a dual verify communication system DSP/BIOS is built in the platform, synchronization and coordination of DSP data processing and ARM system control are achieved, DSP end-to-target detection and tracking algorithms are encapsulated and the underlying algorithm library is genereted, algorithms is optimized combining with characteristics of their own development platform hardware to meet the real-time requirements in the practical application of the system.(3) In this paper, Linux multi thread technology is used in the development of a video surveillance system, which ensures that the video acquiring, output can be synchronized to achieve during operation of the system.Experiments and test results have shown that all modules of the system are functioning properly, it is a rather efficient daily traffic environment monitoring program,can achieve the detection and tracking for traffic moving vehicle satisfactorily,achieve the requirements of timeliness and accuracy in the actual application...