Algorithm Research On Video Moving Object

The video tracking system has many excellent advantages of visual scene, high resistance of electromagnetic interference and good cost performance. As one of its key projects, video moving object tracking syncretizes advanced technologies in many fields such as image processing, pattern recognition, artificial intelligence, automatic control and computer science. Because the results of object tacking contain a great deal of space-time information of the scene, the video tracking technology has got great development since 1960s. It has been applied in many fields such as military vision guidance, security monitoring, traffic control and medical diagnosis.This paper researches on video moving object tacking algorithms. Two common algorithms histogram equalization and unsharp method are discussed at first and then image wavelet transform theory is introduced. In order to enhace infrared image, the method of using platform histogram equalization and non-linear gain function wavelet transform to enhance the global and local image is proposed. Compared to the two algorithms mentioned above, better visual result can be obtained by experiment. In the process of video moving object extraction the common methods such as image different, optical flow and snake method are given. After discussing their merits and drawbacks, an object-based method is present to extract object in real scene. The result indicates that the proposed method can get the continuous edge contour and the real object at the same time. Finally, this paper uses the Kalman filter to track moving object and obtain the tracking curve.