Research On Visual Servo Tracking System Of Small Airborne Self-stabilizing Pan-tilt

As a research hotspot in recent years,visual object tracking technology has broad application space and research value in the field of UAV and visual servo technology.The image information acquired by airborne camera of the UAV is fed back to the UAV system to control the UAV to achieve the goal of tracking the target.Usually UAVs are complex to operate and have a large workload when carrying additional equipment.Therefore,as a bridge between UAV and camera,it is necessary to stabilize the optical axis of the camera to obtain clear image information while independently undertaking real-time target tracking.Aiming at the quadrotor UAV system,this paper presents a low-cost Pan-tilt servo control system based on visual object tracking.The system uses an airborne camera as a visual feedback device to input image information into the host computer.While tracking the target through the visual target tracking algorithm,the position information of the target in the pixel plane is obtained,and the pixel errors between the target position and the center of the image plane are also obtained and input into the servo controller.The tracking effects of three digital servo controllers,linear controller,incremental PID controller and linear quadratic regulator(LQR),are designed and compared to ensure the real-time and stable control of the rotating optical axis of the camera,to reduce the error between the target point and the center point of the image plane,and achieve the visual servo tracking effect of the platform.At the same time,the tracking effects of the two visual object tracking algorithms are compared,and the advantages and disadvantages of each tracking algorithm are analyzed.In order to improve the self-stability of the Pan-tilt and ensure the stability of the camera to obtain image information,an improved gradient descent method based on low-cost three-axis brushless platform is proposed to solve the problem of limited computing power of the Pan-tilt.The algorithm uses quaternion to describe the system model,uses Kalman filter to preprocess the output data of accelerometer,so as to reduce the dynamic error to obtain the optimal attitude which can adapt to complex environment.The real-time stability of airborne camera is guaranteed while the computation is small.By showing that in two groups of experiments,the improved algorithm with gradient descent algorithm has higher accuracy,static and dynamic performance were increased by 15.25% and 27.53%.At the same time,compared with explicit complementary filtering algorithm,it proves the feasibility of the improved algorithm in the Pan-tilt system.