四旋翼无人机单目视觉自主着陆系统研究

On the monocular vision-based autonomous landing system of the unmanned quadrotor helicopter

以基于计算机视觉技术实现四旋翼无人机自主着陆为目标,采用Pixhawk开源飞控和mavros技术,设计黑白相间的正方形分级嵌套地标,然后通过单目摄像头采集图像数据输入机载图像处理上位机NVIDIA TX2中进行图像处理,并结合OpenCV函数库功能完成地标检测识别,最终实现四旋翼无人机位姿解算与自主着陆.

This paper starts from the introduction to the navigation technology of the unmanned quad-rotor helicopter, takes the computer-based visual technology to realize the autonomous landing of the unmanned quad-rotor helicopter as the objective, designs a square cascade-nested landmark chequered with black and white by adopting the Pixhawk open-sourced flight control and the mavros technology, then collects image data and inputs into the airborne upper computer NVIDIA TX2 for image processing, and completes the landmark detection and identification relying on the OpenCV function library, thus realizing the pose position and orientation calculation as well as autonomous landing of the unmanned quad-rotor helicopter finally.