Research On Moving Target Tracking Method And System Design

Moving target tracking technology is an important topic in computer field which is widely used in medical diagnosis,intelligent robot,human-computer interaction,education entertainment,etc.Associated with computer hardware and software technology development is rapid,makes the computer's processing speed faster and cheaper.Domestic and foreign scholars also began to attach importance to the moving target tracking technology,have many effective theory and tracking algorithm is proposed.However,due to the complexity of the scene,the change of illumination,the influence of the occlusion and deformation of the moving target,there are still problems to be solved in theory and application.Therefore,it is also a challenge to study the tracking method of better robustness,accuracy and real-time application in real-time application scenarios.It also has important theoretical significance and practical value.This paper mainly studies and implements the pretreatment process of moving target.Pretreating the image first by weight average method,the median filter,the minimum square method,Canny algorithm,and the mixed Gaussian model.Then,by Kalman filter,we can predict the location of the moving target at the next sampling moment,reduce the target search scope,and improve the tracking efficiency.Then focus on the research and implementation of SIFT,Mean Shift algorithm.After Analyzing their strengths and weaknesses,the MSO algorithm of Mean Shift is proposed.It solves the problem that Mean Shift is insensitive to scale change and the target movement is too fast,which can lead to the problem of tracking effect.In order to meet the requirement of real time and accuracy,it is necessary to meet the practical situation.The experimental results show that MSO algorithm has a significant improvement over Mean Shift algorithm,higher accuracy and shorter time,more suitable for higher real-time requirements than Mean Shift and SIFT algorithm,which can be more accurate and real-time in the system of higher real-time requirements.Moving target tracking system using USB camera to collect images,using the rotation of the camera haeundae implementation,using Linux as a development platform,for the construction of the embedded environment,and the Boot Loader,kernel,root file and Opencv for transplantation,writes the PC desktop applications to realize image display.Finally,the overall test is conducted and the system basically can achieve the expected goal.