| 摘要: |
| 无人机空中加油是一种能有效提升巡航里程及续航时长的技术手段，近距相对位置和姿态测量技术是其中需要解决的关键问题之一。针对该问题研究了无人机自主空中对接中的视觉导航方法，完成了近距对接的地面实验。首先，利用移动对接图标和无人机的GPS/INS信息进行无人机的粗略导航，完成会合；再充分利用视觉图标的颜色与形状信息，通过颜色分割选取目标可能区域，在这些区域中进行快速椭圆检测，针对椭圆检测算法存在误检测及边缘不重合的问题，提出了椭圆检测与轮廓检测相结合的方法，能够更准确地描述图标边缘；最后，利用改进的OI算法进行相对位姿的估计，实现近距的精确导航。实验结果表明，无人机在较高速度下的跟踪效果良好，采用的视觉导航方法能够满足空中对接中精度与实时性的要求。 |
| 关键词: 无人机 空中对接 视觉导航 位姿估计 椭圆检测 |
| DOI： |
|
| 基金项目:装备预研教育部联合基金（6141A02022340） |
|
| Visual Navigation Method in Drone Aerial Docking |
| LIU Ai-chao,SHE Hao-ping,YANG Qin-ning,ZHOU Si-cheng |
| (School of Aerospace Engineering, Beijing Institute of Technology, Beijing 100081, China) |
| Abstract: |
| UAV air refueling is a technical means to effectively improve the range and duration of the cruising. Close-range relative position and attitude measurement technology is one of the key issues that need to be solved. In this paper, a visual navigation method in the autonomous air docking of drones is studied, and the ground experiment of close docking is performed. Firstly, the mobile docking icon and the GPS/INS information of the drone are used for rough navigation of the drone to complete the meeting; then the color and shape of the visual icon is fully utilized, target possible areas are selected by color segmentation. In these areas, fast ellipse detection is carried out. In ellipse detection algorithm, there are problems of false detection and edge misalignment. A method combining ellipse detection and contour detection is proposed. Finally, an improved OI algorithm is used to estimate the relative pose and attitude. The experimental results show that the tracking effect of the UAV is good at higher speeds, and the visual navigation method can meet the requirements of accuracy and real-time in the drone air docking. |
| Key words: Drone Aerial docking Visual navigation Pose estimation Ellipse detection |
