Local binary pattern based ghost elimination in objects detection

We address the problem of ghost elimination for moving object detection in highway surveillance videos. The proposed method is based on LBP (Local Binary Pattern) feature and it is efficient to eliminate ghosts in background modelling-based object detection. The proposed method consists of three steps. The first one is the extraction of the candidate moving objects which have the ghosts. The second one is used to calculate LBP feature value from each pixel in the candidate moving objects. In the final step we achieve ghost elimination by comparing LBP feature value of each pixel with the threshold selected from experiment. The experimental results show that this method is efficient to eliminate the ghosts in background modelling-based object detection.