Research Of Motion Detection And Tracking Based On Video Sequences

Vision is a significant way to acquire information and recognize the world, especially the dynamic visual information, occupies most of the composition in the environment information perceived by human beings.Therefore, the dynamic visual information has become an important research field of the computer vision. The technology of moving objects detection and tracking based on video sequences makes full use of the dynamic correlation among images, and becomes a major subject which has interested many researchers in recent years.This paper has finished a deep theoretic study and sufficient practical works in object detection and tracking from the static background in video sequences:On the research of moving object detection:some common approaches in the field of moving objects detection are studied, these approaches include temporal differencing and background subtraction. Through the characteristics analysis of the two approaches, the paper implements the Surendra estimation to update the background in the background subtraction and the five-frame differencing.Concerned about the advantages of these two methods, it proposes a new method that combines adaptive background subtraction and five-frame differencing,and the experimental result shows that this method can detect the object better.On the research of moving object tracking:A preliminary investigation is made for moving objects tracking. It introduces Kalman filter and Mean shift algorithm, and carries out experiments with the two methods separately. According to the advantages of the two methods, it designs the method that Mean shift algorithm based on motion estimation. Motion estimation is to predict the location of the tracked object by means of Kalman filter and ascertain the scope for searching the tracked object; and then search the object in the scope by Mean shift.This method ensures the right track even when the shelter and similar objects interfere, and improves the reliability of tracking.