摘要

In order to surveille multiple passersby in an alleyway, we propose a system with multi-cameras. We bring forward an algorithm for roughly tracking multiple passersby from monocular video sequences based on the slow movement of the passersby. We separate passersby after merging by color code, which is achieved by the binary encoding of continuum-removed color histogram. Finally, we explain how to match people when they walk from one video to another. We build a suppositional scene by 3dmax, and text our system on it. The experimental results prove that our method is efficient.