基于AprilTag二维码的无人机着陆引导方法

计算机视觉导航技术具有精度高、不受电子干扰,成本低等特点,被誉为未来无人机自主定位的必备手段之一.作为新的定位手段,我们需要在实验室搭建仿真平台来进行大量的模拟训练.将虚拟现实、可视化技术与计算视觉导航算法紧密结合起来,开发了一种基于虚拟仿真环境的无人机视觉自主定位仿真验证系统.在场景中构建虚拟摄像机、棋盘格以及合作二维码标签,在虚拟场景中进行相机标定,并用标定得到的内参解算得到基于标签定位的位置参数,实时计算和输出位置参数.实验结果表明:该仿真系统可以对场景中虚拟相机进行准确标定得到虚拟摄像机的内参,解算定位参数误差在0.2m以内,满足无人机自主飞行以及降落的精度要求,并具备良好的用户显控界面,验证了基于视觉无人机自主定位算法的可行性.

Method of UAV Position Based on Cooperative Two-Dimension Code

Computer vision navigation technology has the characteristics of high accuracy,no electronic interference,and low cost.It is hailed as one of the necessary means for autonomous positioning of future drones.As a new positioning method,we need to build a simulation platform in the laboratory for a large number of simulation training.Combining virtual reality,visualization technology and computer vision navigation algorithms,a UAV vision autonomous positioning simulation verification system based on virtual simulation environment was developed.Construct virtual cameras,checkerboards,and cooperative QR code tags,then,the internal parameters obtained from the calibration are used to solve the position parameters based on label positioning,and the position parameters are calculated and output in real time.The experimental results show that the simulation system can accurately calibrate the virtual camera in the scene to obtain the internal parameters of the virtual camera,and the error of the positioning parameter is within 0.2 meters.It meets the accuracy requirements for autonomous flight and landing of the drone,and has a good user display and control interface,which verifies the feasibility of the autonomous positioning algorithm based on visual drone.