摘要

Visual tracking, when handling the appearance of features in video monitoring, is a challenging problem. In this paper, inspired by the histogram method for counting people with wide-angle camera, we have improved this method for widely used video surveillance with processing speed and background updating. In our improved method, the histogram expression and the target recognition method have been improved. In particular, we have introduced a low-level object descriptor for target classification and a histogram method for target tracking. Also, our improved histogram method could not only be used in indoor spaces but also in out-door spaces. The experimental results presented in this paper show that our method tracks the object accurately and reliably in realistic videos.

全文