摘 要:鉴于机器视觉导航方式具有安全可靠,导航精度高等优点,文章提出一种基于机器视觉的植保无人机自主着陆算法,并对该算法进行了研究:采用自动阈值的 Canny 算子对着陆地面图标进行边缘检测和目标识别,采用亚像素级 Harris 角点检测算法对着陆地面标志图像进行特征提取和特征匹配,根据特征点匹配估计出无人机相对着陆点地面标志的三维位姿。并采用仿真软件对算法进行了对比实验,验证了所用算法的有效性和准确性。
关键词:机器视觉;植保无人机;着陆标志识别;位姿估计;自主着陆
DOI:10.19850/j.cnki.2096-4706.2021.11.013
基金项目:2018 年度湖南省自然科学基金 科教联合基金项目“基于机器视觉的农业植保无 人机仿地自主飞行技术研究”(2018JJ5061)
中图分类号:TP391.4 文献标识码:A 文章编号:2096-4706(2021)11-0048-04
Research on Autonomous Landing Technology of Spraying Drone Based on Machine Vision
YU Kunlin, XIE Zhiming, LIU Jianshan
(Changsha Aeronautical Vocational and Technical College, Changsha 410124, China)
Abstract: In view of the advantages of safe, reliable and high navigation accuracy of machine vision navigation mode, an autonomous landing algorithm of spraying drone based on machine vision is proposed and studied in this paper: Canny operator with automatic threshold is used for edge detection and target recognition of landing ground icons, the sub-pixel Harris corner detection algorithm is used to extract and match the features of the landing ground sign images, and the three-dimensional pose of the spraying drone relative to the ground sign of the landing point is estimated according to the feature point matching. The simulation software is used to compare the algorithms, and the effectiveness and accuracy of the used algorithms are verified.
Keywords: machine vision; spraying drone; landing sign recognition; pose estimation; autonomous landing
参考文献:
[1] 刘全攀,王正杰,王寰 . 基于双目视觉 - 惯性导航的轻型 无人机导航算法 [J]. 兵工学报,2020,41(S2):241-248.
[2] 徐刚,徐耀彬,王小强,等 . 基于机器视觉的四旋翼导航 平台的设计与实现 [J]. 电子制作,2018(1):23-25.
[3] 汪烈兵,姜雄飞,石春光,等 . 基于图像滤波与 Hough 变换的红外弱小目标检测 [J]. 红外技术,2020,42(7):683- 687.
[4] 李静,陈桂芬,丁小奇 . 基于改进 Canny 算法的图像边 缘检测方法研究 [J]. 计算机仿真,2021,38(4):371-375.
[5] 孙青锋.角点检测的舰船图像配准算法 [J].舰船科学技术, 2021,43(12):193-195.
[6] 孟学斌 . 基于机器视觉的旋翼无人机自主着陆系统设计与 实现 [D]. 呼和浩特:内蒙古工业大学,2019.
作者简介:于坤林(1975—),男,汉族,湖北广水人,教授, 硕士,研究方向:图像处理。