Classification and segmentation approach for detecting moving object in road applications

In this paper, we bring up the problem of moving object detection which is very essential for many applications in computer vision. In fact, so many algorithms are proposed in the research field these last years. The basic idea of our paper is to propose an efficient segmentation and classification approach based on a discriminating classifier capable of distinguishing between moving objects and static objects in real time. Thus, static objects are considered to belong to the background and moving objects will be surrounded by a bounding box facilitating their tracking. In fact our proposed algorithm proves itself by the experimental results which show its strength by having the highest rate of recognition and localization precision comparing to other classical methods.