Computer Engineering and Applications ›› 2022, Vol. 58 ›› Issue (17): 306-313.DOI: 10.3778/j.issn.1002-8331.2101-0333

• Engineering and Applications • Previous Articles     Next Articles

Landmark-Based Lightweight UAV Accurate Landing Algorithm

YUAN Hui, PEI Chu, WANG Shuai, LI Jinsong, JIANG Min, TONG Jiapeng   

  1. 1.Power Research Institute, State Grid Shanxi Electric Company, Taiyuan 030001, China
    2.Department of Automation, North China Electric Power University, Baoding, Hebei 071003, China
  • Online:2022-09-01 Published:2022-09-01



  1. 1.国网山西省电力公司 电力科学研究院,太原 030001 
    2.华北电力大学 自动化系,河北 保定 071003

Abstract: The on-board computer of lightweight UAVs is usually weak in computing performance, which is difficult to meet the needs of accurate landing. To solve this problem, a landmark-based lightweight and accurate landing algorithm is proposed. The landing logo can be detected quickly and in real time by identifying contrast colors and specified shapes. Image processing is simple rapid and accurate. Then by calculating the relative position of UAV and landing landmark in the two-dimensional plane, drone can get a landing location and direction. The algorithm execution process does not need to consider the camera focal length. The actual test results show that in some specific application scenarios, such as UAV autonomous charging, this algorithm has a simple process, higher stability and landing accuracy than the traditional method.

Key words: unmanned aerial vehicle(UAV), image recognition, relative position, autonomous precision landing, landing system

摘要: 轻量化无人机的机载计算机通常计算性能较弱,很难满足无人机实现精准降落的需要。针对这一问题,提出了一种基于地标的轻量化精准降落算法,通过识别对比颜色和指定形状实现快速实时地检测着陆标识,图像处理流程简单快速且准确,通过相对位置计算在二维层面得到无人机对于降落地标的相对位置和方向,引导无人机精准降落,算法执行过程不需要考虑相机焦距。实际测试结果表明,在一些特定如无人机自主充电的应用场景,该算法过程简单,稳定性和降落精度相较于传统方法较高。

关键词: 无人机, 图像识别, 相对位置, 自主精准降落, 降落系统