Research On The Technologies Of Visual Tracking

Visual tracking is a new arising research task in recent years, which integrates technologies of computer vision, pattern recognition, artificial intelligence and so on. It has a lot of potential applications in security surveillance, intelligent traffic system, video compression and index, etc.Image sequences are the input of visual tracking system, and the system returns some attributes (size, position, velocity and so on) of the object. Ideally, all the information should be outputted in real-time and accurately. But in real world, it is very hard to get the ideal output because of noises. In this paper, we present a visual tracking system, which can resist noises in real world. First, several common algorithms in visual object detection are introduced-frame difference, background subtraction, optical flow and active contour. The first two methods are simple in principle and easy to implement, so they have a lot of applications. The other two methods have better performances in background noises and moving background, but they are complicated and comptutationally intensive. Then we will introduce some common-used filtering techniques. Particle filter which is based on Bayesian estimation and Monte Carlo method will be emphasized. On the basis of last two chapters, a new algorithm, whose frame is particle filter, is proposed. The color information and motion information will be fused in our algorithm, which can improve the robustness and the accuracy. Besides that, we will discuss shadow in videos. Shadow has great impact on visual tracking system, especially in some applications that need to measure the size of the object. It is a meaningful job to deal with shadow.