交通场景中的实时多目标跟踪方法

摘要

针对多目标检测与跟踪技术的应用要求,提出一种用于交通场景的实时多目标跟踪方法.设计背景建模算法和基于角点动态特征的分层跟踪算法,利用背景建模算法提取视频帧前景,并在前景区域运用分层跟踪算法,包括Comer、Cluster和object 3层架构,通过在不同层之间引入对应的聚类算法,以实现目标跟踪.实验结果表明,该方法适用于多数交通场景,对光照、阴影具有较强的鲁棒性.%To address the issue that more and more requirements to the multi-objects detection and tracking technology, this paper proposes a real time multi-objects tracking algorithm in traffic scene. It designs background modeling algorithm and the dynamic multi-level tracking algorithm based on comers. Background modeling algorithm is used to extract the foreground. Tracking algorithm has three levels which are Corner, Cluster and Object It introduces suitable grouping algorithm to the different layers to realize object tracking. Experimental results show that the proposed method can be applied to most traffic scenes. It is robust to light and shadow and to the multi-objects tracking with complex background.