Moving shadows detection algorithm and implementation based on dual background modeling

Moving object images in video sequence are obtained by subtracting the background images with the current frame images, background extraction is an important process. Firstly, in this paper, the background and moving objects are extracted by using the improved Gaussian mixture background modeling. Then, the previous moving object and moving object of frame difference are effectively integrated into a moving foreground object we are interested in. Finally, we extract the background region under the moving foreground object. Due to the influence of illumination, the foreground objects contain moving shadows, which affect the performance of moving object detection. This paper proposes a simple and effective method which uses the difference of texture feature and color information in gray foreground image and gray background image to remove the shadow of the foreground object, and then detect the real foreground object.