机器视觉辅助的插头锥套式无人机自主空中加油仿真  

Machine Vision Aided Simulation of Probe and Drogue Unmanned Aerial Vehicle Autonomous Aerial Refueling

作　　者：王旭峰[1] 董新民[1] 孔星炜[1] 

机构地区：[1]空军工程大学航空航天工程学院,西安710038

出　　处：《科学技术与工程》2013年第18期5245-5250,共6页Science Technology and Engineering

摘　　要：为准确获取无人机自主空中加油对接阶段受油插头与加油锥套的相对位姿信息,提出一种机器视觉辅助的插头锥套式无人机自主空中加油方案。研究了机器视觉识别跟踪加油锥套的算法,利用卡尔曼滤波算法估计无人机与加油锥套的相对位姿。实验结果表明:机器视觉图像处理算法可精确识别跟踪加油锥套,滤波器估计的相对位姿误差收敛速度较快,满足插头锥套式无人机自主空中加油的需要。In order to precisely obtain the relative pose between probe and drogue during unmanned aerial vehicle（UAV） autonomous aerial refueling docking,a machine vision aided simulation scheme of probe-drogue UAV autonomous aerial refueling is proposed.Based on the machine vision,the recognition and tracking algorithm of refueling drogue is investigated.The relative pose between UAV and refueling drogue is estimated by using of Kalman filter algorithm.Experiment results show that the machine vision aided image processing algorithm can recognize and track refueling drogue precisely and the convergence of relative pose errors estimated by the filter algorithm is proved,fulfilling the demand of probe-drogue UAV autonomous aerial refueling.

关 键 词：无人机 自主空中加油 机器视觉 锥套识别跟踪 相对位姿估计 

分 类 号：V249.1[航空宇航科学与技术—飞行器设计]