基于视频的智能交通监控系统技术研究

【作者】 徐德文；

【导师】 史忠科；

【作者基本信息】 西北工业大学 ， 控制理论与控制工程， 2004， 硕士

【摘要】 交通视频监控是智能交通系统的重要组成部分，然而目前普遍使用的交通视频监控系统是采用PC作为系统的视频图像处理平台，不可避免的存在稳定性差、实时性低等问题。为了弥补这些不足，论文采用高性能的DSPs取代传统的PC作为核心处理器，设计了一种基于DSPs的交通视频监控系统的图像处理平台，详细的对其硬件实现和软件调试进行了论述；在这个基础上分析了视频图像采集和处理的流程，讨论了一种利用EDMA的实时后台传输数字视频方法和三帧缓冲的存储器管理方法，为系统的实时性提供了保障。 论文结合教研室在车流量和车速检测方面的经验，在这个硬件平台上，采用一种基于检测线的快速运动检测算法对车流量检测进行了仿真验证。最后对系统的技术实现和应用领域等方面进行了展望。更多还原

【Abstract】 Video Surveillance is one of the important parts of Intelligent Transportation System(ITS). Nowadays, Popular Computers(PC) are often used as central control unit in image processing in most Video Surveillance systems, however, many serious problems such as instability and low processing speed are existed in the those systems. In order to deal with those problems above, a video surveillance scheme based on DSPs was presented, which used high performance DSPs to replace PC as central control unit. Not only did we realize the hardware circuits of the system, but also we developed it’s software. Furthermore, three essential methods, for instance the data transfer method by Extend Direct Memory Access(EDMA), buffer management method by Three Buffer Strategy and optimized data acquire method had been discussed.Based on the experiences of our teaching and research room in intelligent traffic monitoring, a real-time motion detection algorithm is used in our hardware system to detect multiple traffic parameters such as the number of the vehicles and the average speed and so on. Experiment results show that the performance of the system is satisfied. Finally, the potential application fields of the system are proposed.更多还原