The Research Of Target Tracking Control Based On Visual In UAV

In recent years,with the high-speed development of commercial Unmanned Aerial Vehicles(UAV),accelerated the application research of multirotor.There are many relevant research directions in the field of UAV's target tracking control,Such as multi-source information fusion,UAV's vision target recognition,UAV's flight control,etc.This paper focuses on the research of quadcopter's visual tracking control.Apply the method of visual control to achieve visual recognition and flight tracking for goal through quadcopter.Locating the target by visual algorithm and planning path based on the positioning information,controlling the quadcopter to move along with the desired direction.Under the Robot Operating System(ROS),applying visual algorithm and path planning algorithm based on probability are combined to achieve goal tracking through UAV.First of all,under the condition of fixed flight altitude,the image sequence of target is obtained by the visual sensor,and the Trackinglearning-detection(TLD)is adopted to achieve image recognition of moving target.According to the position of the identified moving target in the current image,the relative position of the target and the UAV is calculated,the GPS information of the UAV is combined with the relative position information of the target,and the position of the target in the world coordinate system is estimated.Then the path planning algorithm based on probability is used to plan the moving path,and the UAV is controlled to track the target along the planned path.Aiming at the problem of visual tracking control of UAV,the visual tracking,path planning and flight control based on quadcopter are studied.The effects of different visual tracking algorithms and path planning algorithm on UAV flight tracking are studied,and the corresponding simulation experiments are given.The research background,the research significance and visual control on UAV at home and abroad are analyzed.The motion control,mathematical modeling and controller design of UAV are studied,and the simulation experiments are carried out.The single target visual tracking algorithm of UAV is studied.The principle of KCF and TLD algorithm are studied,and the experimental results of the two algorithms are analyzed.It is found that the recognition accuracy of TLD algorithm is much higher than that of KCF algorithm in the case of target loss and target recognition.The TLD algorithm is considered to be suitable for the visual tracking control of UAV.Then the path planning algorithm of UAV is studied,and the path planning algorithms A* and RRT* are studied respectively.The simulation results show that the RRT* path planning algorithm performs better than A* algorithm in the adaptive ability of environment change and the real-time performance of the algorithm.It is considered that RRT* path planning algorithm is suitable for UAV path planning.Based on the above research,the experimental scheme of UAV tracking control system is presented.The system is tested and the result data are analyzed.Finally,the main contents of the research and the problems encountered in the course of the research are summarized,and suggestions for the future research direction of the project are given.