Research On Uav Target Tracking Based On Machine Vision

With the progress and development of science and technology,in the field of machine vision tracking technology and video images of the moving object and more attention based on machine vision,target tracking of UAV platform is one of the hot areas of concern.Because of its small size,light weight,easy to operate.vertical take-off and landing and hovering in the air,multi rotor unmanned aerial vehicle(UAV)has been widely used in all aspects of daily life,especially in the field of video image capture.The target tracking technolog is used to target UAV platform on the video image,background environment and moving target’s interference and occlusion,can effective real-time tracking is the main content of this thesis.This thesis in the first to detect moving targets,the commonly used detection methods of background environment under static condition,inter frame difference method,optical flow method,background difference method,and the related experiments to make comparison and analysis of the above method,finally choose the continuous frame difference method and background difference method combined with two kinds of calculation methods were small and fast running time.to use the algorithm to improve the accuracy of moving,target detection.Secondly,in the part of the moving target tracking,target tracking algorithm commonly used Mean-shift algorithm,Cam-shift algorithm and particle filtering method for analysis on the principle and specific process of algorithm in visual tracking.According to the advantages and disadvantages of the above methods,this thesis combines the particle filter algorithm and the Mean-shift mean vector offset algorithm to improve the particle filter method.Finally,through the use of drones for moving targets in the background environment has different interference and occlusion cases.using Mean-shift algorithm,particle filter algorithm and the improved Cam-shift method were carried out in the video of the imoving target tracking experiment,after analyzed after dray in the background color without blocking interference when the Cam-shift algorithm runs fast efficiency there is high,the particle filter algorithm and occlusion of target and background colors are similar when using the modified,better target tracking.