A Novel Approach based on Optical Flow for Moving Object Detection

摘要

Each application that benefit from smart video processing has different needs, thus requires different treatment. However, they have something in common: moving objects. We address the problem of detection and tracking of moving objects in a video stream obtained from a moving camera. In this paper, a novel approach for detection of moving object based on optical flow is compared with the existing block based method. Existing block-based method uses global motion parameters to distinguish moving objects from the background wherein novel approach discussed in this paper includes two principal steps, the first step removes the camera motion using block matching algorithms and second step aims at detection of the motion of objects based on optical flow. The motion between two frames is commonly described by a dense displacement vector field which links the location of each point in a given frame to its location in the next frame. Optical flow gives a description of this motion.