航空学报 > 2009, Vol. 30 Issue (8): 1503-1507

基于序列图像的无人机自测速方法与试验

于起峰,李强,雷志辉,尚洋,朱宪伟,刘晓春   

  1. 国防科学技术大学 航天与材料工程学院
  • 收稿日期:2008-06-17 修回日期:2008-11-05 出版日期:2009-08-25 发布日期:2009-08-25
  • 通讯作者: 于起峰

Methods and Experiments of Velocity Measurement of UAV from Image Sequence

Yu Qifeng, Li Qiang, Lei Zhihui, Shang Yang, Zhu Xianwei, Liu Xiaochun   

  1. College of Aerospace and Materials Engineering, National University of  Defense Technology
  • Received:2008-06-17 Revised:2008-11-05 Online:2009-08-25 Published:2009-08-25
  • Contact: Yu Qifeng

摘要: 为自主、精确测量无人机的飞行速度,研究并发展了一种基于序列图像的无人机自测速方法并进行了相应的试验。该方法利用无人机上已有的惯导装置、高度计和摄像机,在连续成像的条件下,通过匹配跟踪得到地面同名点在相邻两帧光学或红外实时图中的位置,利用飞行高度、姿态信息和成像帧频计算得到无人机的瞬时飞行速度。在无人机的匀速平飞段,通过大量数据拟合得到高精度的平均飞行速度。通过挂飞试验对方法进行了验证,实时得到了小于0.2 m/s的测速结果,满足工程要求的精度,为工程应用打下了一定的基础。

关键词: 视觉导航, 自测速, 序列图像, 匹配跟踪

Abstract: To measure the velocity of an unmanned aerial vehicle (UAV) independently and accurately, a method based on image sequences is presented and corresponding experiments are carried out. After matching and tracking ground features among the sequential images given by the visible or infrared camera, the method uses the frame rate and the flying parameters provided by the inertial sensors and laser rangefinders to obtain the instantaneous velocity. While the UAV is flying in uniform velocity, the average velocity is received by data processing. Flight test are performed, from which accurate results better than 0.2 m/s are received in real time. The experimental results can meet the requirements of the engineering standard, and thus lay a certain foundation for engineering application.

Key words: vision navigation, velocity measurement, image sequence, matching and tracking

中图分类号: