Graph Optimization based Visual SLAM fusing KeyPoints and Markers

Zhen Chen, Yang Zhou, Fengdi Zhang, Min Xu, Xiangdong Liu, Zhen Li

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

1 Citation (Scopus)

Abstract

Simultaneous localization and mapping (SLAM) plays an important role in autonomous navigation for mobile robots. Most of the visual SLAM approaches use keypoints for tracking, whose performance however suffers from the unstable landmarks during task due to uncertain light condition and frequently changeable viewpoint. The situation even becomes worse for visual SLAM in low texture environment especially in indoor buildings, where the supplementary artificial markers can be used to improve the robust detection under a wider range of circumstances. Inspired by this thought, this paper developed a visual SLAM system integrated with keypoints and artificial markers. A graph optimization problem is constructed to optimize the trajectory by taking both the reprojection error of keypoints and the influence of markers into consideration. The experimental results on SPM dataset demonstrate the superior accuracy of the proposed graph optimization algorithm compared with the start-of-the-art ORB-SLAM2.

Original languageEnglish
Title of host publicationProceedings of the 39th Chinese Control Conference, CCC 2020
EditorsJun Fu, Jian Sun
PublisherIEEE Computer Society
Pages3483-3488
Number of pages6
ISBN (Electronic)9789881563903
DOIs
Publication statusPublished - Jul 2020
Event39th Chinese Control Conference, CCC 2020 - Shenyang, China
Duration: 27 Jul 202029 Jul 2020

Publication series

NameChinese Control Conference, CCC
Volume2020-July
ISSN (Print)1934-1768
ISSN (Electronic)2161-2927

Conference

Conference39th Chinese Control Conference, CCC 2020
Country/TerritoryChina
CityShenyang
Period27/07/2029/07/20

Keywords

  • Visual simultaneous localization and mapping (SLAM)
  • artificial markers
  • graph optimization
  • keypoints

Fingerprint

Dive into the research topics of 'Graph Optimization based Visual SLAM fusing KeyPoints and Markers'. Together they form a unique fingerprint.

Cite this