A Survey of Motion Detection in Image Sequences

Detection of motion is an essential and important step in many applications such as in the use of video surveliance in security systems, anomaly detection, traffic monitoring, people tracking. Detection of the moving object is a growing research area because of its extensive applications in real life. This paper emphasizes on survey of different algorithms for detection of motion. The most commonly used strategy of subtraction of background is described in detail. A survey of different techniques of motion detection is presented in this paper. The novel method of detection of motion for the pedestrian (indoor) without the need of finger printing of location proposed by Zohuis explained in detail. The advantages of schemes presented are also mentioned. The novel technique of detecting both global and local motion which is coherent, proposed by Yunpeng Wu is described in detail with its advantages and comparison with the previous related works. The applications of detection of motion are presented in this paper. A detailed description of widely used datasets along with the examples of frames of the datasets is given in this paper.