Research on Background Reconstruction Based on the Judging of Model Categories

The key factor of background difference method is to reach accurate background reconstruction from video image sequence. In this paper, we present an algorithm of background reconstruction approach based on model categories judging. First we realize the conversion from RGB color space to HSV color space, and establish the initial background model categories according to the pixels and its statistical characteristics of HSV color space. Second we update the background model categories dynamically, add new background model categories. Then we detect the moving objects by judging the same pixels in follow-up video information with the known categories, and detect and remove the shadow of objects. Experiment shows that this approach has the advantage of effective.