Target Recognition And Tracking Control For UAV Based On The Vision

The cooperation of UAV(Unmanned Aerial Vehicles)and USV(Unmanned Surface Vehicles)has become a hotspot of the study of multiple robots system at home and broad.Not only for it can extend the radius of flight and range of missions of UAV,but also can provide an aerial perspective and early warning surveillance.The accurate positioning and tracking control of UAV to USV,is one of the important research content.Based on the situation that the UAV know the GPS informations of the USV’s,this thesis studies emphatically the special target detecting and recognizing tied on the landing pad.By using the visual processing algorithm and the position tracking control algorithm,this thesis finishs the accurate positioning and tracking control for the moving target.Major research contents include:Firstly,the image processing algorithm based on ArUco(Augmented reality utility compiler)code is used to realize the detection and recognition of the specific target by image preprocessing,candidate tag search,perspective transformation and tag decoding.The accurate position of the target in the visual image is given for the UAV dynamic tracking to the moving target.Secondly,this thesis chooses the DJI M100 four rotorcraft as the test platform,and focuses on the attitude control system of the UAV.Using the PID control algorithm,this thesis designs the UAV position tracking control system based on the visual iamge recognition.Finally,the UAV vision-based experimental system is finished in this thesis and has been described from both hardware aspect and software aspect.Several experiments have been done to test the specific target and the tracking control.The results and data of the experiments show the effectiveness of the visual recognition and tracking control system,which can provide an experimental method for the accurate positioning and tracking control of the UAV.