Research On Target Extraction And Tracking Algorithm Based On Optical Flow Method

Target tracking can be widely applied in our life and also very important in industry.To achieve the goal of locate position of targets in the tracking process,there are some difficulties we have to overcome,such as changes of gesture,dynamic background and block problem.This paper focuses attention to applying optical flow to the tracking framework.Optical flow is a very important algorithm in tracking and also a method of solving the moving parameters in the image by using the optical flow equation.Due to the lack of constraints,many optical-flow methods based on other assumptions are proposed.So based on the Pyramid LK optical flow method,this paper has some deep research,and finds some points which can be improved and puts forward the solution strategy to improve the performance.The main work is as follows:(1)The first part introduces the definition and basic framework of optical flow tracking,and also the theory behind the optical flow equation.It can be divided into two kinds of methods: dense optical flow and sparse optical flow.This paper adopts the sparse optical flow which relies on feature points matching.The principle and deficiency of feature points and optical flow are introduced.And then,this paper introduces two commonly used optical flow methods,comparing the advantages and disadvantages of the two optical flow methods,explaining the necessity of further optimizing the correlation algorithm.(2)The second part proposes a new method based on sparse optical flow to address the problem of target extraction and tracking in dynamic backgrounds.First,the pyramid of the LK optical flow method is used to generate an optical flow image to match the feature points between two images.Second,the feature points are divided preliminarily based on the optical flow information on the displacement and direction of the optical flow image.In the case of blocked targets in subsequent frames,we apply the Kalman estimation method to predict the target location and locate the target quickly upon its reappearance.The proposed method demonstrates excellent performance in meeting real-time requirements in dynamic backgrounds and can be applied to tracking slow-or fast-moving targets in blocked or unobstructed scenes.(3)For the problem of similar interferences and partial occlusion in target tracking,the third part proposes a new method based on sparse optical flow and SVT.First,optical-flow equation and SVT is combined to establish over-determined equations.And then Kalman algorithm is introduced to estimate the motion parameter of target to ensure the score of trainer in partial occlusion.The experiment shows the proposed algorithm can avoid interferences of similar vehicle and improve the tracking accuracy in partial occlusion.