An illumination-robust algorithm based on visual background extractor for moving object detection
Huan Wang, Qinglin Wang, Yuan Li, Yan Liu, Yaping Dai · 2015
Aiming at the problem that ViBe (Visual Background extractor) algorithm is sensitive to illumination variation, an improved ViBe algorithm is proposed in this paper. First, a brightness reference image is build, which represents the brightness level of background. Second, adjust the brightness of current frame, which is changed by illumination variation, back to the reference images brightness level adaptively by using image intensity normalization method based on histogram statistics. With this method, the influence of illumination variation can be basically eliminated. At last, a morphology postprocessing is used to filter the results of binary image, remove the noise in the background, fill small voids within the object, and suppress the influence of dynamic background. The experimental results show that compared with several existing algorithms, the proposed algorithm has higher detection accuracy and is robust for illumination variation.