基于单目视觉的路面车辆检测及跟踪方法综述

首先介绍了车辆检测算法的3种基本组成部分：检测、验证、跟踪，然后根据算法的组成重点介绍了车辆检测以及跟踪的几种主要算法。车辆检测算法包括基于特征的方法、基于光流场的方法和基于模型的方法，车辆跟踪算法包括基于区域相关的方法、基于活动轮廓的方法、基于特征的方法和MeanShifi快速跟踪算法。根据试验结果对各种车辆检测和跟踪方法的优点、缺点以及实际应用中不同情况下适用范围的局限性进行了综合分析。最后在结论部分总结展望了文中介绍的几种车辆检测和跟踪方法的应用前景，并提出了在实际应用时的一些建议和将来的主要研究和发展方向。

Vehicle Detection and Tracking Based on Monocular Vision

First, the three component of the vehicle detection algorithm including detection, verification and tracking are discussed.Then, the algorithm of detection and tracking are discussed with emphasis on composition.The vehicle detection algorithm includes feature based, optical flow based and model based method, The vehicle tracking algorithm includes region correlation based, active contours based, feature based and mean shift based method.The merit and disadvantage of these algorithms is discussed according to the result of experimentation, Finally, some suggestions for future research and application are presented.