基于视觉的小型四旋翼飞行器跟踪系统研究

Tracking System for Small Quadrotor UAV Based on Vision

针对小型四旋翼飞行器跟踪地面移动目标的问题,提出一种基于视觉的地面移动目标跟踪方法并搭建了基于Linux的飞行实验平台。在跟踪过程中,由搭载的云台相机采集图像数据并使用Open CV库函数对图像数据进行实时处理;通过改进的压缩跟踪算法对图像数据中的目标进行实时的目标跟踪;通过建立的地面移动目标运动模型计算目标的位置状态信息;由ROS平台通过坐标系转换估计四旋翼飞行器运动数据并向飞行器的飞行控制系统发送具体的飞行指令。最后,通过具体的四旋翼飞行器目标跟踪实验来验证此方法的可行性。

Aiming at the problem of small quadrotor UAV tracking ground targets, a vision based ground moving target tracking system is proposed and a linux-based flight experiment platform is built. In the process of tracking, image data collected by the Zenmuse X3 camera and the Open CV library function is used to process the image data in real time;The target in image data is tracked by the improved compression tracking algorithm; The position status information of the target is calculated by establishing the ground moving target motion model; The ROS platform is used to estimate the motion data of the quadrotors and send specific flight instructions to the flight control system of the aircraft. Finally, the feasibility of this method is verified by the specific four-rotor aircraft target tracking experiment.