A Visual Localization System for Complex Indoor Environment

X. L. Jiang, Chao Yang Liu, Hongbin Deng*

*此作品的通讯作者

科研成果: 书/报告/会议事项章节会议稿件同行评审

摘要

In recent years, unmanned technology has been widely used in military and civil fields. However, agents often stop working because of positioning problems in complex urban environments. In this paper, we present a real-time Visual Localization System for complex indoor scenes. we describe the problem of state estimation in visual localization. The essence of it is to obtain the maximum likelihood estimation of the state variables from the camera observation data. In our System, we use ORB feature to estimate the motion of the camera between adjacent frames after balancing the accuracy and efficiency of system. We also maintain a local map in the system and use nonlinear optimization to further optimize the estimation results. Through the test of TUM RGB-D benchmark, the system can operate in real time with 21FPS. The results show that the triaxial position error is about 0.03 m, and the attitude error is 0.06 rad.

源语言英语
主期刊名Proceedings of 2022 International Conference on Autonomous Unmanned Systems, ICAUS 2022
编辑Wenxing Fu, Mancang Gu, Yifeng Niu
出版商Springer Science and Business Media Deutschland GmbH
1482-1492
页数11
ISBN(印刷版)9789819904785
DOI
出版状态已出版 - 2023
活动International Conference on Autonomous Unmanned Systems, ICAUS 2022 - Xi'an, 中国
期限: 23 9月 202225 9月 2022

出版系列

姓名Lecture Notes in Electrical Engineering
1010 LNEE
ISSN(印刷版)1876-1100
ISSN(电子版)1876-1119

会议

会议International Conference on Autonomous Unmanned Systems, ICAUS 2022
国家/地区中国
Xi'an
时期23/09/2225/09/22

指纹

探究 'A Visual Localization System for Complex Indoor Environment' 的科研主题。它们共同构成独一无二的指纹。

引用此