基于视觉的煤矿输电线巡检无人机自主降落系统  

Vision Based Autonomous Landing System of UAV for Coal Mine Transmission Line Inspection

作　　者：张传江[1] 刘险峰 陈帮干 曹金芳[2] 孟祥宇 ZHANG Chuanjiang;LIU Xianfeng;CHEN Banggan;CAO Jinfang;MENG Xiangyu(Huaibei Mining(Group)Co.,Ltd.,Huaibei 235000,China;Huaibei Mining Culture and Tourism Media Co.,Ltd.,Huaibei 235000,China;Xutuan Mine of Huaibei Mining Co.,Ltd.,Huaibei 235000,China)

机构地区：[1]淮北矿业(集团)有限责任公司机电装备处,安徽淮北235000 [2]淮北矿业(集团)有限责任公司,安徽淮北235000 [3]淮北矿业文化旅游传媒有限公司,安徽淮北235000 [4]淮北矿业股份有限公司许疃矿,安徽淮北235000

出　　处：《现代信息科技》2020年第4期14-16,共3页Modern Information Technology

摘　　要：面对煤矿中复杂的地形,巡检无人机很难将降落地标与环境区分开,无法实现定点的精准降落。基于以上问题,提出了一种基于视觉导航的煤矿输电线无人机自主降落系统。为了区分地标和环境区,将地标设计为一种拓扑模式,为了提高图像二值化的检测效率,提出了一种动态阈值方法。根据有效图像信息计算水平面内的相对距离,用线性插值法求出相对高度。在相应平台上进行降落实验,实验结果表明,提出的自主降落系统性能稳定,可以准确自主降落。Facing the complex terrain in the coal mine,it is difficult for UAV to distinguish the landing landmark from the environment and realize the precise landing at a fixed point.Based on the above problems,an autonomous landing system of UAV for coal mine transmission line based on visual navigation is proposed.In order to distinguish the landmark from the environment area,the landmark is designed as a topological mode.In order to improve the detection efficiency of two value image,a dynamic threshold method is proposed.According to the effective image information,the relative distance in the horizontal plane is calculated,and the relative height is calculated by linear interpolation.The landing experiment is carried out on the corresponding platform.The experimental results show that the proposed autonomous landing system has stable performance and can land accurately.

关 键 词：计算机视觉 图像二值化 无人机 自主降落 

分 类 号：V279[航空宇航科学与技术—飞行器设计] V249