现代电子技术2025，Vol.48Issue(9)：154-160,7.DOI:10.16652/j.issn.1004-373x.2025.09.023
基于SLAM与深度学习融合的室内无人机系统
Indoor UAV system based on fusion of SLAM and deep learning
摘要
Abstract
In view of the limited positioning accuracy,imperfections of the design and weak robustness of the positioning system in the current indoor UAV design,a design scheme for indoor UAV(unmanned aerial vehicle)system based on the fusion of LiDAR SLAM(simultaneous localization and mapping)and deep learning is proposed.In this scheme,a miniaturized design is adopted for the UAV.LiDAR and binocular camera T265 are selected for indoor mapping and navigation,the main control system of the UAV is equipped with a program error protection mechanism to enhance the robustness of the UAV effectively,the indoor mapping and positioning are completed by the Cartographer algorithm,and the FastestDet deep learning algorithm is integrated to complete relocation and image correction,so as to achieve high-precision indoor navigation flight.The experimental results show that the proposed UAV system can realize indoor autonomous mapping and navigation flight according to the set path and task.The system has fast operation speed,high positioning accuracy,and strong operation stability,so it has high practical value.关键词
无人机/室内自主飞行/激光雷达/SLAM/Cartographer/FastestDetKey words
UAV/indoor autonomous flight/LiDAR/SLAM/Cartographer/FastestDet分类
电子信息工程引用本文复制引用
易磊,黄哲玮,陆韬宇..基于SLAM与深度学习融合的室内无人机系统[J].现代电子技术,2025,48(9):154-160,7.基金项目
教育部2020年第二批新工科研究与实践项目:为未来而教,面向"电气化+",重构电气工程专业卓越领军人才培养体系(E-NYDQHGC20202219) （E-NYDQHGC20202219）
