Research On Tracking Four Rotor Aircraft Target

Four-rotor UAV generally has three motion modes: pitch,roll and yaw,which can realize the basic functions of vertical takeoff and landing,fixed-point hovering,and can adapt to forward flight,side flight and reverse flight postures according to its basic motion mode.In addition,the UAV with flight control has the ability of autonomous takeoff,and can realize corresponding functions according to the program,so as to adapt more.Different environments.This kind of UAV has been widely used in many fields,such as farmland detection,road monitoring,military reconnaissance and so on.In this paper,Four-rotor UAV is taken as the research object,and the visual tracking system of this UAV is deeply studied and designed.In the process of research,two kinds of visual tracking algorithms are adopted.The UAV end and ground station end are designed,and the design is implemented on the physical platform.Firstly,this paper takes the Four-rotor UAV as the research object,and designs the experimental platform with KCF target tracking algorithm.The experimental platform consists of the following parts: open source small four-axis,UAV platform,WIFItransmissionmodule,5.8G analog map transmission,Win10+QT5+Opencv3.4.0.The system target tracking algorithm is implemented in the ground station,and the ground station is developed with QT+Opencv3.4.0.Secondly,the KCF target tracking algorithm is introduced for the image sequence collected by UAV.This algorithm is based on Kalman filter.By analyzing the test results on the test set,the interference caused by some occlusion and light changes can be effectively avoided.However,the shortcomings and shortcomings of KCF are also summarized through testing,mainly in the process of aerial photography if the object moves quickly or the camera shakes too much,resulting in poor shooting effect.In order to improve this situation,the prediction and correction function is applied to KCF algorithm.Finally,the ground station control system is designed based on OpenCV library.In order to obtain the pixel coordinates,the real coordinates are effectively converted through the analysis of camera coordinate system,pixel coordinate system and image coordinate system.Through the vision platform of the miniature UAV developed by the punctual atom,the system has completed the active tracking of the target.