基于深度相机和二维码的室内移动机器人定位技术
Indoor mobile robot localization technology based on depth camera and QR code
唐振宇 1张兆威 2蒋林2
作者信息
- 1. 武汉科技大学冶金装备及其控制教育部重点实验室,湖北 武汉 430081
- 2. 武汉科技大学冶金装备及其控制教育部重点实验室,湖北 武汉 430081;武汉科技大学机器人与智能系统研究院,湖北 武汉 430081
- 折叠
摘要
提出一种以ORB_SLAM2 为基本框架、使用二维码辅助深度相机的室内移动机器人定位方法.针对现有的视觉SLAM在定位过程中出现的Z轴漂移现象,提出平面运动模型约束,以降低机器人定位结果中的Z轴累计误差;针对视觉SLAM在弱纹理环境中算法退化、容易跟踪失败等问题,在室内环境中张贴二维码并将其数学模型作为定位约束,以提高系统的准确性与鲁棒性.真实环境下的实验结果表明:相比原始算法,所提算法在Z轴精度上提高了42.95%;以ORB_SLAM3的定位结果为真值,所提算法的定位精度提高了4.11%,该算法在室内环境下具有定位优势.
Abstract
A method for indoor mobile robot localization was proposed based on ORB_SLAM2 as a basic framework and using QR code assisted depth cameras.In response to the Z-axis drift phenomenon that occurred in the positioning process of existing visual SLAM,a planar motion model constraint was proposed to reduce the cumulative Z-axis error in robot positioning results.In response to the problems of algorithm degradation and easy tracking failure in weak texture environments of visual SLAM,QR codes in indoor environments were posted and their mathematical models were used as localization constraints to improve the accuracy and robustness of the system.The experimental results in real environments showed that compared to the original algorithm,our algorithm had improved the Z-axis accuracy by 42.95%;Using the positioning result of ORB_SLAM3 in real environment,the positioning accuracy of our algorithm was improved by 4.11%compared to the original algorithm,proving the positioning advantage of our algorithm in indoor environments.
关键词
移动机器人/同步定位/地图构建/深度相机/二维码/平面运动模型约束Key words
mobile robots/synchronous positioning/map construction/depth camera/QR code/constraints for planar motion models引用本文复制引用
出版年
2024
国家科技期刊平台
NSTL
万方数据