吴文秀 关志全 汪磊 郑军斌 阳璞琼

摘 要：无人机广泛应用于军事、农业、交通运输等方面，具有巨大的应用潜力和广阔前景。受电池能量密度和能量利用效率等因素的制约，无人机机载电池没有足够的续航能力，发挥不了最大优势。为提高无人机的飞行续航能力，采用无线充电基站对无人机进行充电。为达到无线充电的最大充电效率，需要实现无人机精确降落。为了实现无人机的定点降落，提高定点飞行能力，提出了一种GPS和视觉识别相结合定点降落方法。首先通过pixhawk飞控模块配合M8N模块来进行GPS定位，使无人机到达充电基站附近，然后通过再启用视觉辅助（openmv）实现较精准的降落。实验结果表明，无人机能够准确识别充电基站从而实现定点降落。

关键词： 无人机;视觉识别;openmv;定点降落

文章编号： 2095-2163（2021）01-0041-05 中图分类号：V279;V249 文献标志码：A

【Abstract】Unmanned Aerial Vehicle （UAV） is widely used in military， agriculture， transportation and other fields， and has great application potential and broad prospects. Restricted by factors such as battery energy density and energy utilization efficiency， the onboard battery of UAV does not have enough endurance， which can't give full play to its advantages. In order to improve the flight endurance of UAV， wireless charging base station is used to charge UAV. In order to achieve the maximum charging efficiency of wireless charging， it is necessary to realize the accurate landing of UAV. In order to realize the fixed-point landing of UAV and improve its fixed-point flight capability， a fixed-point landing method combining GPS and visual recognition is proposed. At first， the pixhawk flight control module is matched with M8N module to carry out GPS positioning， so that the UAV can reach the vicinity of the charging base station， and then， the visual aid （openmv） is enabled again to achieve a more accurate landing. The experimental results show that the UAV can accurately identify the charging base station and realize fixed-point landing.

【Key words】UAV; visual recognition; openmv; fixed point landing

0 引 言

目前，各類科技创新有效地推动了现代无人机技术的发展，无人机广泛用于气象探测、灾害监测、农药喷洒、地质勘测、地图测绘、交通管制等诸多领域[1-5]。近些年来，无人机技术取得了长足进步，且应用领域正在不断拓展，对其各项技术指标也随即提出了更高的要求。传统的无人机技术主要依靠惯性导航系统（Inertial Navigation System， INS）和全球定位系统（Global Position System， GPS）进行导航，然而，导航过程中的惯性器件具有累积误差，对初始值过于敏感，而GPS并不总是可获取的，并且即使是可获取，精度也无法满足无人机导航要求[6-7]。随着现代技术的发展，计算机视觉导航开始逐步走上无人机技术领域的舞台[8-11]。……