Abstract
In recent years, monocular simultaneous localization and mapping (SLAM) algorithms have achieved good performance, while most of them could not deal with dynamic environments. This paper proposes a novel monocular SLAM system with mask loop closing (MLC-SLAM) to deal with dynamic objects. The proposed system combines the direct and feature-based methods to track fast and robustly. The proposed approach achieves motion refinement and loop closing in the feature-based module, which exploits semantic information by training a visual vocabulary offline with semantic labels from the mask network. Evaluation on the TUM mono VO dataset shows that MLC-SLAM achieves better performance than the state-of-the-art monocular SLAM algorithms: DSO and ORB-SLAM, in terms of accuracy, robustness, and the performance of loop closing.
Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
