The Recognition And Tracking Of Moving Objects Based On Characteristic

With the continuous development of computer vision technology, moving object tracking is more and more worth to study. It plays an important role in traffic control, perimeter protection, and intelligent tracking, et al.We are going to focus on intelligent tracking. In this paper, we try to track the teacher who is teaching in a classroom.We will use the background subtraction method to extract the moving points from a video. We merge the moving points which are close to each other to form moving regions.Shadow detection is also an important problem in video processing. We will discuss this problem and give an algorithm based on previous work. It is traditional and effective to solve the problem of image segmentation by constructing certain energy function in the framework of Markov random fields. Shadow detection can be seen as segmenting an image into two parts:"shadow"and"non-shadow". So we can use the framework of Markov random fields to solve the problem of shadow detection. How to define the energy function is the key issue in the Markov random field framework. This paper will give one way to define the energy function, which can do better in the problem of shadow detection.We use the feature matching method to track the moving object that we are interested in when there are more than one moving objects.