基于道路的运动车辆检测

基于图像的运动车辆检测是智能交通系统中一个重要而又困难的问题，它通常受到光照、路边摇晃树木、以及各种恶劣天气等的影响。提出了一种基于道路的车辆检测方法，在道路的入口处只利用背景差，对非入口处运动背景和不同道路上运动车辆的运动向量分别进行高斯建模，并利用贝叶斯准则和多帧信息来检测出运动车辆和运动背景。实验结果显示该方法取得了很好的效果。

Detection of moving vehicle road-based

Image- based moving vehicle detection is an important yet difficult problem in ITS. It is usually influenced by illumination, waving trees, abnormal weather and so on. This paper proposes road - based detection of moving vehicle. In the entry of road, the moving vehicle is segmented only by background subtraction. While in other region, the motion vector of dynamic background and moving vehicle in different roads by Gaussian motion model is modeled. And the dynamic background and moving vehicle by Bayesian rule and in formation of multi - frame are detected. The result of is segmented experiment demonstrates that the method proposed has a good effect.