Moving Objects Detection In Complex Background

The moving objects detection in video sequence using background subtractionapproach is studied in this thesis. Firstly a new background extraction approach basedon moving objects point accumulation is proposed. The approach takes advantage offewer moving object pixels to accumulate moving objects point in the video sequencein stead of background pixels to update the background image. Secondly the shadowfilter is adopted to remove shadow. Based on the knowledge of the gray-scale value ofthe shadow is less than background and moving objects. The experimental results showthat the algorithm improves the speed of background extraction, and gets bettercontour of the moving objects.