The issue of detecting the status of video scenes was studied. First, Gaussian pyramid algorithm was employed to preprocess input frames. By comparing three consecutive inter-frame difference methods, the method that calculated the inter-frame difference on the gray frame images was adopted. According to the spatial distribution of the difference information as well as the morphological features of high intensity points, a new algorithm named Ratio of Erosion and Dilation per Point (REDP) was proposed to detect status change under different videos, which were captured in different scenes, luminance and weather conditions. The experimental results confirmed the effectiveness of indicating whether there are motions in the video. When the proposed approach was applied to the inter-frames, the detection capability of the algorithm can become more robust under an appropriate frame interval.
