Kalman filtering based object tracking in surveillance video system

In the field of motion estimation for surveillance video, various techniques have been applied. One of the common approaches is Kalman filtering technique and it is interesting to explore the extension of this technique for the prediction and estimation of motion via the image sequences. In this paper, a moving object tracking in surveillance video using Kalman filter is proposed. The typical Kalman filter is good in tracking the position of a moving object. However, when dealing with occlusion, the typical Kalman filter is not able to keep tracking and predicting the position of the occluded moving object. During occlusion, the information of moving object is not available for detection and tracking. The lacking of occlusion scene determination and prediction ability cause the existing Kalman filter fails in tracking occluded object. Besides that, in the case of tracking multiple moving objects, existing Kalman filter will experience difficulties to identify the respective objects. Therefore, in order to encounter these problems, an object tracking method using enhanced Kalman filter will be developed. The ability of tracking occluded moving object will be added to increase the efficiency during tracking. Furthermore, object recognition feature will be added too to increase the accuracy of the object tracking system.