| Digital image processing and computer vision are developing very fast in these years, and they are applied popularly in civil and military use. Automatic target tracking technology is one of the hot topics of this area. According to the needs of the project, this paper studies the methods of detecting and tracking moving targets in video images, builds an experimental platform, and realizes these methods on the platform to simulate the automatic tracking equipment.Firstly, for problems of moving target detecting, this paper studies the frame difference method and the background difference method. It gives the principle and the process in detail, and has made the contrastive experiment. Analysising the experiment results, it concludes the advantages and disadvantages of the two methods, and gives the application area of them.Secondly, for target tracking, this paper studies the tracking method based-on Kalman motion estimation and the CamShift tracking method. The tracking method based-on Kalman motion estimation uses the motion information for tracking; The CamShift tracking method uses the color image instead of gray image which is used popularly by traditional methods. The CamShift tracking method adopts the character information of the target, uses the color histogram of the target as character, the back-project image of the histogram as the matching template to match and search for realizing the target tracking. This paper detailedly gives the principle and process of the two tracking methods, makes experiments. It points out the application area of each other.Finally, proper detecting and tracking methods for the project are selected, software is made with VC++6.0. To validating the methods and software, an experimental platform of tracking system is built. This paper presents the constitution, the functions, and the operation process of the platform detailedly. A tracking experiment is done in the real scene with the platform, and the experiment results are given. |
PDF Full Text Request