Research On Visual Target Tracking Algorithm In Complex Scene

With the deeply rapid development of modern computer technology,visual target tracking technology has become one of the most important issues in the current research.Because of the increasing level of image processing technology,the accuracy rate of pattern recognition has been greatly improved,and with the development of the theory of probability and statistics,improvement of the computer in the hardware,it is very worth studying.Visual target tracking combined with knowledge of image processing,image recognition and image content understanding is mainly based on acquiring multiple detection targets from images,obtaining the state of detecting image targets,and automatically tracking objects.After research and use,target tracking technology has been widely used in many fields,but the problems that it shows can not be ignored,for example,the performance of the tracking algorithm is easily affected by the environment of the target.At present,the visual target tracking in complex scene is still more complex than other visual target tracking scenes.The visual target tracking is difficult in complex scene,which is influenced by illumination intensity and background and so on,making the description of the target feature more difficult.At the same time,the moving state of the target in complex scenes has nonlinear characteristics,which puts forward higher requirements for the general tracking algorithm.In this paper,on the basis of particle filtering,the characteristics of color space,direction and SIFT are studied.A target tracking algorithm based on HSV features and data association is proposed.Finally,the experimental results show that the algorithm meets the expected requirements of complex environment and multi target tracking.