Moving Targets Detection And Tracking Under The Dynamic Scenes

Moving target detection and tracking is very important to the field of computer vision and the basic steps, the field of computer vision is an important branch. Outside world, information is transmitted through a variety of sensory organs in the human brain, which, in addition to the visual sensory system other than the information obtained combined together only account for about 20%, this is not unimportant that other sensory systems On the contrary, not only the outside world of visual information that large, but also further illustrates the importance of the visual system of the special nature. With the development of computer technology, people's dream is to give people the function of the machine, especially the eye part. People try to use the computer to realize the visual information processing, gradually to develop a new discipline, that is computer vision computer vision, human eyes instead of using the machine to be stationary or moving object detection, tracking, judgments, etc., the main research is to use a computer to simulate the human visual system. Generation of computer vision began in the 20th century,60s, and 80s made a major breakthrough. Fully integrated computer vision is a highly technical, it is characterized not only the function of the human eye in the simulation, it is sometimes important to complete the human eye can not complete some of the work computer vision in the digital image processing developed on the basis of the new discipline in computer science and technology, signal processing, cognitive science, physics, statistics and applied mathematics, neurophysiology research fields, in the military, medical diagnostic, manufacturing, document analysis, safety testing, civil, transportation and other areas of the various intelligence applications, has very broad prospects. It can be said, moving target detection and tracking of this technology has covered all aspects of society.Dynamic context of moving target detection and tracking of the main task is to get the camera to process the video image sequence, in each frame of video images, accurate and complete to clearly separate the foreground object region; calculates each frame on the target location, In the video stream on a continuous moving target position, and so provide the target trajectory.Based on the moving target detection, tracking, digital image processing, global motion estimation, cam shift algorithm learning and research, mainly to complete the following work:First, the introduction to this field of research significance and status at home and abroad, and introduced the static background of several common methods of detection and dynamic detection of the context of several common methods, these methods of their unique advantages and shortcomings, there is more applicable environment.Again, briefly introduced the basics of digital image and to the noise of the method and mathematical morphology methods, including dilation and erosion, opening closing operation and other operations. Also cited several common basis for tracking methods. Discusses their advantages and disadvantages and so on.Third, the introduction of the principle of global motion estimation and compensation, block matching principle and the method to deal with the case of dynamic background, combined with the frame difference method to detect moving target objects.Finally, Kalman filtering cam shift mean shift combination of these methods to achieve tracking of moving target objects. This approach can significantly reduce the computation, and enhance the system in real time.As science and technology development and social progress, moving target detection and tracking technology will increasingly improve, I hope this work can be done moving target detection and tracking technology make a modest contribution.