Complex Scenes The Real-time Tracking Of Moving Targets

Along with a rapid development of hardware techniques, there has been a growing interest in the area of analyzing dynamic images. Especially, tracking a moving object has become a very important research topic since it can be used in many applications. The difficult problem in visual tracking is performing fast and reliable matching of target from frame to frame. This paper describes a real-time tracking system which tracked moving object by controlling a servo device in such a way that make the target always lie at the center of image frame.Histogram character is not affected by object figure's variance and scale, and it has good stability. Mean shift algorithm is an applied algorithm that calculates local optimal result fast and effectively. Therefore, the kernel-tracking algorithm in our system is based on the mode of histogram and the search method of Mean Shift algorithm. Supplemented by the motion verifying of candidate target, our system runs fast and effectively during tracking processing.We also utilize a Kalman filler to encode a tracking history into state parameter of filter. The estimated state parameter will be used to reduce search areas for model matching and to control a servo device.Besides in order to erase the influence of camera moving, we applied a approach to registry image by matching characteristic points in the image. Its main idea is to search for the corresponding characteristic points between base image and following image, and also project the following image to base image by computing the transformation between images with LMV. Based on these methods, some steps such as characteristic points' matching are improved. At the same time, we add the validation of characteristic points to assure it robust when doing transform.The real time tracking on moving object is realized with combing the three modules, which are object model matching, object motion estimation and servo controlling processing.