ORB-SLAM2: An Open-Source SLAM System for Monocular, Stereo, and RGB-D Cameras

• Published in: IEEE transactions on robotics Vol. 33; no. 5; pp. 1255 - 1262
• Main Authors: Mur-Artal, Raul, Tardos, Juan D.
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.10.2017; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Bundle adjustment; Cameras; Central processing units; CPUs; Driving; Feature extraction; Localization; mapping; Odometers; Optimization; RGB-D; Simultaneous localization and mapping; simultaneous localization and mapping (SLAM); Source code; State of the art; stereo; Tracking loops; Trajectory; Trajectory analysis
• ISSN: 1552-3098; 1941-0468
• DOI: 10.1109/TRO.2017.2705103

We present ORB-SLAM2, a complete simultaneous localization and mapping (SLAM) system for monocular, stereo and RGB-D cameras, including map reuse, loop closing, and relocalization capabilities. The system works in real time on standard central processing units in a wide variety of environments from small hand-held indoors sequences, to drones flying in industrial environments and cars driving around a city. Our back-end, based on bundle adjustment with monocular and stereo observations, allows for accurate trajectory estimation with metric scale. Our system includes a lightweight localization mode that leverages visual odometry tracks for unmapped regions and matches with map points that allow for zero-drift localization. The evaluation on 29 popular public sequences shows that our method achieves state-of-the-art accuracy, being in most cases the most accurate SLAM solution. We publish the source code, not only for the benefit of the SLAM community, but with the aim of being an out-of-the-box SLAM solution for researchers in other fields.