Target Tracking Technology Research Based On Multiple Algorithm Fusion

Target tracking technology, especially on the human body moving target tracking, such as in computer vision technology,it is a fusion image graphics, pattern recognition, computer technology and simulation, artificial intelligence, human-computer engineering technologies and it is the most dynamic branch. In national security, public security detection, traffic regulation, and other fields play an important role increasingly raised.In this paper, based on a lot of moving target tracking technology related references carefully study and analysis, based on the theory of density estimation, using the basic form of mean shift algorithm, I do a thorough research and analysis, studies the parameter density estimates the density estimation with no parameter and the basic form of mean shift, also make the proof. When obstacles appear, introducing the similarity function, use the kalman filter, to put forward a solution. Finally, feasibility experiment to verify it. To teace face,use detail characteristics, matrix into an adaptive enhancement algorithm, and the continuous adaptive mean shift algorithm, improve the real-time tracking, to realize the of face detection. By setting the feedback function, each frame detection frequency and increase the accuracy and success of face tracking. Experimental tests show that on the basis of the human body has successful track, through the adaptive window size, to achieve the goal tracking face again.This paper studies the moving target tracking method needs to be further improved in real time, in order to used in the defense, security, civil and other fields get more extensive application.