针对智能交通对分布式智能视频监控的需求,设计了基于DM6446嵌入式系统的视频车辆检测跟踪系统。提出一套简单有效的车辆检测跟踪算法,包括自适应背景更新算法、基于边缘检测和最优双阈值分割融合的运动区域检测、改进的快速连通区域判断方法以及基于多特征的跟踪算法。同时,对DM6446平台软硬件及其开发方法进行介绍,并针对该平台介绍了如何移植和代码优化。最终通过在DM6446平台实际运行表明,该系统能实时有效地检测和跟踪车辆,具有较高准确率和稳健性。
In order to make distributed intelligent video surveillance system meet the intelligence transportation' s needs, video vehicle detection-tracking embedded system based on DM6446 is designed. A simple and effective vehicle detection tracking algorithm is proposed, including the adaptive background updating algorithm, motion regions detection based on the integration of the edge detection and the optimal threshold, the improved rapid method of identifying the connected regions, and the tracking algorithm based on multi-feature. Meanwhile, the hardware and software of DM6446 platform and development methods are introduced, and how to transplant and optimize the codes in this platform is described. Finally, the actual running through the DM6446 platform shows that the system can effectively detect and track with high accuracy and robustness.