摘要
针对室内视觉定位特征稀疏时鲁棒性较低的问题,文章利用惯导元件输出频率大、短期精度高的特点,提出了一种低负载的单目视觉和惯性测量单元(inertial measurement unit, IMU)信息融合方法。该方法在初始化阶段根据位移相似性原理设计算法解算视觉位移尺度信息,在导航阶段将视觉信息输出的状态量通过尺度旋转变换后作为观测量进行卡尔曼滤波。通过实验验证了该方法的有效性,实验结果证明,所设计的视觉惯导组合导航系统在提升处理效率的同时,全过程均方根误差比单一同步定位与制图(simultaneous localization and mapping,SLAM)系统下降了17.1%。
Aiming at the problem of low robustness when indoor visual positioning features are sparse, combined with the features of large output frequency and high short-term precision of the inertial navigation element, a low-load monocular vision and inertial measurement unit(IMU) information fusion method is proposed. In the initialization stage, the similarity principle is used to design the algorithm to solve the visual displacement scale information. In the navigation stage, the state of the visual information output is subjected to Kalman filtering as an observation measurement after being transformed by the scale factor and the rotation matrix. The effectiveness of the algorithm is verified by experiments. The experimental results show that the designed integrated navigation system improves the processing efficiency while the root mean square error(RMSE) of the whole process is 17.1% lower than that of the single simultaneous localization and mapping(SLAM) system.
作者
王睿
陈国良
许晓东
曹晓祥
张超
WANG Rui;CHEN Guoliang;XU Xiaodong;CAO Xiaoxiang;ZHANG Chao(NASG Key Laboratory of Land Environment and Disaster Monitoring,China University of Mining and Technology,Xuzhou 221116,China;School of Environment Science and Spatial Informatics,China University of Mining and Technology,Xuzhou 221116,China)
出处
《合肥工业大学学报（自然科学版）》
CAS
北大核心
2019年第8期1077-1082,共6页
Journal of Hefei University of Technology：Natural Science
基金
国家重点研发计划资助项目(2016YFB0502105)
江苏省自然科学基金资助项目(BK20161181)
江苏高校品牌专业建设工程资助项目(PPZY2015B144)
关键词
视觉定位
惯性导航
组合导航
尺度信息
卡尔曼滤波
visual positioning
inertial navigation
syncretic navigation
scale factor
Kalman filtering
