基于云台相机的四旋翼无人机跟踪控制系统

针对四旋翼无人机跟拍过程中视角固定，易丢失目标的问题，设计了一种基于云台相机的四旋翼无人机跟踪控制系统。首先，使用云台相机对地面移动目标进行拍摄，通过对目标的颜色特征和形状特征进行检测识别，使用核相关滤波( KCF: Kernel Correlation Filter) 方法进行视觉跟踪，得到地面移动目标在图像坐标系的位置。然后，通过建立针孔模型，解算出在云台相机带有俯仰角时的无人机与地面目标的相对位置关系。最后，设计了离散串级比例-积分-微分( PID: Proportion-Integral-Differential) 踪控制器，实现对无人机的位移控制，使四旋翼无人机可对地面移动目标进行稳定跟踪。同时设计了串级比例-微分( PD: Proportion-Differential) 控制器以实现云台相机的视角跟踪，增大了无人机的跟踪范围，降低了丢失目标的风险。实验结果表明，所设计的四旋翼无人机跟踪控制系统可实现对地面移动目标的稳定跟踪。

Quadrotor Tracking Control System Based on PTZ Camera

Aiming at the problem that the quadrotor has fined angle of view during the tracking process and is easy to lose the targot，a four-rotor UAV ( Unmanned Aerial Vehicle) tracking control system based on pan /tilt camera is designed for the problem of quadrotor UAV． Firstly，the head camera is used to shoot the ground moving target． By detecting and recognizing the color and shape features of the target，the KCF ( Kernel Correlation Filter) method is used for visual tracking，and the position of the ground moving target in the image coordinate system is obtained and used． The method of cascading PD( Proportion-Differential) control is used to move the perspective of the PTZ ( Pan /Tilt /Zoom ) camera to achieve tracking of the target． Then，by establishing a pinhole model，the relative positional relationship between the drone and the ground target when the PTZ camera has a pitch angle is calculated． Finally，a discrete cascading PID ( Proportion-Integral-Differential) controller is designed to realize the tracking of the quadrotor UAV． The experimental results show that the four-rotor UAV tracking control system designed in this paper can achieve stable tracking of ground moving targets．