An updating algorithm of self-adaptive background based on energy method

Abstract

With the large computation quantity, long computing time and low real-time performance, the optical flow is not fit for the complex surroundings. The background image difference is only fit for the still background and so far, no reasonable approach has been designed and implemented for automatic background updating along with the background variance. To solve the above problems, an updating algorithm of self-adaptive background based on energy method (USBE) first is proposed in this paper. In the complicated background, this algorithm is mainly applied to the moving object detection, which can modify and update the coarse background model (CBM) in real time. Moreover, combining the background image difference and energy similarity of moving objects, this algorithm can extract the moving objects completely and its application is universal. Experimental results demonstrate that the proposed updating algorithm of self-adaptive background based on energy method (USBE) can update the variational background exactly and quickly and improve the recognition accuracy of the moving objects.