基于视觉的自动空中加油技术  

Vision-based autonomous aerial refueling

作　　者：慕春棣[1] 李波睿[1] 

机构地区：[1]清华大学自动化系

出　　处：《清华大学学报（自然科学版）》2012年第5期670-676,681,共8页Journal of Tsinghua University(Science and Technology)

摘　　要：无人机在军用和民用领域都发挥着越来越重要的作用,但是续航时间短、负载能力弱等缺点限制了无人机的应用。自动空中加油技术能够增加无人机的续航时间,增强无人机执行长航时、复杂任务的能力,从而有效地弥补上述缺点。计算机视觉技术的发展为无人机自动空中加油技术的实现提供了有效的支持。对基于视觉的硬管式自动空中加油技术的2类典型解决方案进行了介绍和对比,并在此基础上提出了基于切换模式的双目视觉解决方案,总结了涉及的关键技术和待研究的问题。Unmanned aerial vehicles （UAV） are playing more and more important roles in both the military and civilian domains. However, the drawbacks of short duration and low payload capacity significantly limit UAV applications which could involve complex, long term missions. Autonomous aerial refueling （AAR） can effectively resolve some of these limitations by extending the UAV＇s payload and range. Computer vision can support for autonomous aerial refueling. The paper reviews two typical vision-based boom and-receptacle refueling （BRR） approaches and then a switching scheme based stereovision （SSS） that combines the advantages of these two approaches and avoids the disadvantages. The system design and key vision based BRR solutions techniques are described.

关 键 词：导航 无人机 自动空中加油 计算机视觉 位姿估计 多传感器融合 

分 类 号：V249.3[航空宇航科学与技术—飞行器设计]