Research On Maneuvering Target Tracking Algorithm Based On Video Image

The image tracking has already changed the original mode of tchnology in many application fields by providing a new efficient solution. At present, the image tracking is faced with the opportunities for the further development and challenges from various application fields. In this thesis, some approaches to improve the performance of the image tracking algorithm are researched to resolve the problems including the disadvantage in the moving objects detection algorithm and motion parameter estimation.Firstly, the basic method of the vedio image tracking has been discussed in this thesis. Then builded motion parameter estimation algorithm based on “current”statistical model for maneuvering targets and adaptive Kalman filter are proposed through studying the motion dynamics of mobile and analyzing the movement characters of object in image.Then, this thesis have studied the real-time technique of objects tracking based on video image.The “current” statistical model for maneuvering targets and adaptive Kalman filter have been introduced and the techniques of predicting Kalman filtering,adaptive tracking window, adaptive image translation technology, and robust detecting for object pixels are developed to overcome the deficiencies of the threshold method based on frame difference. Last, video image for single non-rigid target and rigid target are researched by covered scenarios under the optimized adaptive tracking window and uncovered scenarios under the adaptive tracking window.After, we have discussed the multi-maneuvering targets tracking theory based on video image and analyzed the defects existing in the traditional track association method.The position and velocity information are introduced as the basic parameters of tracking association method. Then, the method based on motion parameter predicting is proposed, and we have done the theoretical simulation experiments of maneuvering target tracking.Finally, the adaptive tracking window of multiple maneuvering targets based on video image has been put forward in this thesis. the algorithm of objects tracking for multi-maneuvering targets tracking under noise contaminated video image and multi-maneuvering non-rigid targets tracking and covered tracking are selected asengineering background, experiment shows that satisfied performance is achieved by this algorithm.