点击次数：

论文类型：期刊论文

发表时间：2012-09-01

发表刊物：International Journal of Advancements in Computing Technology

收录刊物：EI、Scopus

卷号：4

期号：17

页面范围：326-333

ISSN号：20058039

摘要：In this paper, we propose a novel homography-based multiple cameras tracking algorithm. Firstly, the foreground likelihood map in each view is obtained by using the Codebook background modeling algorithm. Then, the view-to-view homographies induced by the chosen plane are computed by using several static landmarks. Trajectories of multiple objects on the plane are calculated by using the Shortest Path Algorithm (SPA). Finally, we achieve the 3D tracking of multiple objects by combining the trajectories on the selected plane and the results of foreground detection. Compared with other popular methods, our proposed algorithm does not require computing the vanishing points of multiple cameras. Therefore, it reduces the complexity and improves the accuracy in the meanwhile. The tracking efficiency is also improved by using SPA algorithm. The experimental results demonstrate that our method is robust to occlusion and also achieves real-time tracking performance.