Abstract
Traditional vision-based simultaneous localization and mapping (SLAM) technology cannot obtain the semantic information of the surrounding environment, which will cause the robot to fail to complete intelligent grasping, human-computer interaction, and other advanced decision tasks. There are many challenges to solve this problem, such as obtaining less semantic information of the environment with low precision and being unable to deal with dynamic objects in a real environment quickly and effectively. In this paper, a robust visual odometry using semantic information in complex dynamic scenes was designed. Specifically, we present a refined instance segmentation method based on the contextual information of the frame to improve the accuracy of segmentation. On this basis, a feature detection and elimination algorithm for dynamic objects based on instance-level semantic information is proposed to improve the localization accuracy of camera pose. We extensively evaluate our system on the public TUM data set and compare it with ORB-SLAM2 and other methods. Experiments show that our methods greatly improve the localization accuracy of the camera pose and the robustness of the system, which verifies that our system is effective in complex dynamic scenes.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Similar content being viewed by others
References
Garcia-Fidalgo, E., Ortiz, A.: Vision-based topological mapping and localization methods: a survey. Robot. Auton. Syst. 64, 1–20 (2014)
Fuentes-Pacheco, J., Ascencio, J., Rendon-Mancha, J.: Visual simultaneous localization and mapping: a survey. Artif. Intell. Rev. 43, 55–81 (2015)
Klein, G., Murray, D.: Parallel tracking and mapping for small AR workspaces. In: ISMAR, IEEE, pp. 225–234, January, 2007
Mur-Artal, R., Tardos, J.: Orb-slam2: an open-source slam system for monocular, stereo and RGB-D cameras. IEEE Trans. Robot. 33, 1255–1262, October, 2016
Mur-Artal, R., Montiel, J., Tardos, J.: ORB-SLAM: a versatile and accurate monocular slam system. IEEE Trans. Robot. 31, 1147–1163, October, 2015
Rublee, E., Rabaud, V., Konolige, K., et al.: ORB: an efficient alternative to SIFT or SURF. In: International Conference on Computer Vision, IEEE (2012)
Cadena, C., Carlone, L., et al.: Simultaneous localization and mapping: present, future, and the robust-perception age. IEEE Trans. Robot. 32(6), 1309–1332 (2016)
Younes, G., Asmar, D., Shammas, E.: A survey on non-filter-based monocular visual slam systems. Robot. Auton. Syst. 98, 67–88 (2016)
Yu, C., et al.: DS-SLAM: a semantic visual slam towards dynamic environments, pp. 1168–1174, October, 2018
Xiao, L., Wang, J., Qiu, X., Rong, Z., Zou, X.: Dynamic-SLAM: semantic monocular visual localization and mapping based on deep learning in dynamic environment. Robot. Auton. Syst. 117, 1–16 (2019)
He, K., Gkioxari, G., Dollar, P., Girshick, R.: Mask R-CNN. IEEE Trans. Pattern Anal. Mach. Intell. 42, 386–397, June, 2018
Sturm, J., Engelhard, N., et al.: A benchmark for the evaluation of RGB-D SLAM systems. In: 2012 IEEE/RSJ International Conference on Intelligent Robots and Systems, IEEE (2012)
Sun, Y., Liu, M., Meng, M.: Improving RGB-D slam in dynamic environments: a motion removal approach. Robot. Auton. Syst. 89, 110–122 (2016)
Li, S., Lee, D.: RGB-D SLAM in dynamic environments using static point weighting. IEEE Robot. Autom. Lett. 2, 2263–2270, July, 2017
Acknowledgments
The author(s) disclosed receipt of the following financial this article: This work was supported by National Natural Science Foundation of China (No.61872327), Fundamental Research Funds for Central Universities (No. ACAIM190102), Natural Science Foundation of Anhui Province (No. 1708085MF146), the Project of Collaborative Innovation in Anhui Colleges and Universities (Grant No.GXXT-2019-003), the Open Fund of Key Laboratory of Flight Techniques and Flight Safety, (Grant No.2018KF06), Scientific Research Project of Civil Aviation Flight University of China (Grant No.J2020-125) and Open Fund of Key Laboratory of Flight Techniques and Flight Safety, CAAC (Grant No. FZ2020KF02).
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2021 Springer Nature Singapore Pte Ltd.
About this paper
Cite this paper
Wang, H., Wang, L., Fang, B. (2021). Robust Visual Odometry Using Semantic Information in Complex Dynamic Scenes. In: Sun, F., Liu, H., Fang, B. (eds) Cognitive Systems and Signal Processing. ICCSIP 2020. Communications in Computer and Information Science, vol 1397. Springer, Singapore. https://doi.org/10.1007/978-981-16-2336-3_56
Download citation
DOI: https://doi.org/10.1007/978-981-16-2336-3_56
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-16-2335-6
Online ISBN: 978-981-16-2336-3
eBook Packages: Computer ScienceComputer Science (R0)