In this project I have used the concept of frame differencing and background subtraction algorithm to propose a modified algorithm which can be used effectively and accurately with comparison to both frame differencing method and background subtraction model used individually for detecting moving objects in a sequence of frames. In this project we have used the method of frame differencing to propose a new adaptive frame differencing method which shall take account of the velocity of the moving object in order to find the number of frames to be skipped each stage of detection to calculate inter-frame difference in order to get the region of moving object. The above procedure is combined with background subtraction model with a new idea of changing the background dynamically to have a better image of the moving object. The area obtained from adaptive frame differencing is added with the area obtained from adaptive background subtraction model to have a clear view of the pixels associated with the moving object. After getting the detected object the centroid of it is passed to the tracking module in order to track the object in upcoming frames by using the concept of dynamic template matching algorithm which uses a correlation function in order to track the detected object in the region of interest in the upcoming frames. When the tracking fails the algorithm goes back to detection module and the process repeats. Thus we proposed a effective tracking algorithm which can be use even if the object of interest is far away from the camera independent of the motion of the object.
展开▼
