基于轮式机器人与微小目标识别技术的机场跑道异物检测系统研究  

Research on runway foreign object debris detection system based on wheeled robots and small object recognition technology

作　　者：白颢 刘璟之 李雄威 王泽玮 BAI Hao;LIU Jingzhi;LI Xiongwei;WANG Zewei(College of Information Engineering,Hainan Vocational University of Science and Technology,Haikou 571126,China;School of International Education,Nanjing University of the Arts,Nanjing 210013,China;Changzhou Vocational Institute of Engineering,Changzhou 213164,China;Brisight(Hainan)Science and Technology Development Co.,Ltd.,Haikou 571158,China)

机构地区：[1]海南科技职业大学信息工程学院,海南海口571126 [2]南京艺术学院国际教育学院,江苏南京210013 [3]常州工程职业技术学院,江苏常州213164 [4]金景(海南)科技发展有限公司,海南海口571158

出　　处：《无线互联科技》2024年第22期68-72,共5页Wireless Internet Science and Technology

摘　　要：文章针对航空器起降间隙较大的支线机场跑道的异物自动识别与定位问题展开研究,提出了一种基于具备自主导航功能的轮式机器人结合工业相机与激光雷达系统的跑道异物自动识别定位系统。该研究主要采用图像分块处理与深度学习框架相结合的计算机视觉算法,对跑道微小异物目标进行自动识别与定位,具备跑道自主巡航检测的能力。该研究实现了系统样机的集成与跑道环境现场测试的一系列工作,获得了对5 mm直径微小异物80%的综合检出率,有利于该类型机场跑道异物检测的进一步研究并取得了积极的效果。This paper focuses on the issue of automatic detection and localization of foreign objects debris(FOD)on the runway of regional airports,where the intervals between aircraft takeoffs and landings are relatively large.A runway FOD automatic detection and localization system is proposed,based on a wheeled robot equipped with autonomous navigation functions,combined with industrial cameras and a LiDAR system.The research primarily employs a computer vision algorithm that integrates image block processing with a deep learning framework to automatically detect and localize small objects on the runway.Additionally,the system is capable of autonomous runway inspection and navigation.In the study,the integration of the system prototype and a series of on-site tests in the runway environment are successfully conducted,achieving an overall detection rate of 80%for small objects with a diameter of 5 mm,which has the advantage for a further research and positive results for the FOD detection on this type of runway.

关 键 词：轮式机器人 目标识别 机场跑道异物 图像处理 深度学习 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] V351[自动化与计算机技术—计算机科学与技术]