Analysis of computer vision based techniques for motion detection

摘要

Motion detection or moving object detection in a video is a process of identification of the change of physical position of the object. Nowadays the field of computer vision based motion detection has attracted many researchers due to its vast applications in the field of intelligent video surveillance, traffic monitoring, event detection, people tracking and behavior analysis etc. Four basic approaches background subtraction, frame difference; temporal difference and optical flow estimation based motion analysis have been discussed and analyzed in terms of their usability and computational time in this paper. Further the various related work under these approaches have been discussed and analyzed. The review presented in this paper is an attempt to provide the thorough understanding of various traditional approaches for motion detection and to present the current scenario of research in this area.