ORB-SLAM2: An Open-Source SLAM System for Monocular, Stereo, and RGB-D Cameras
ORB-SLAM2: An Open-Source SLAM System for Monocular, Stereo, and RGB-D Cameras
复制标题
DOI：
10.1109/tro.2017.2705103
复制
发表时间：
2017-10-01
影响因子：
7.8
通讯作者：
Tardos, Juan D.
中科院分区：
文献类型：
--
作者：
Mur-Artal, Raul;Tardos, Juan D.
我们提出了ORB-SLAM 2，一个完整的同时定位和映射（SLAM）系统的单目，立体和RGB-D相机，包括地图重用，闭环和重新定位功能。该系统可在各种环境中的标准中央处理器上进行真实的实时工作，从小型手持室内序列到工业环境中的无人机飞行和城市周围的汽车驾驶。我们的后端，基于单目和立体观测的光束平差，允许精确的轨迹估计与度量尺度。我们的系统包括一个轻量级的定位模式，利用视觉里程计跟踪未映射的区域，并与地图点匹配，允许零漂移定位。对29个流行的公共序列的评估表明，我们的方法达到了最先进的精度，在大多数情况下是最准确的SLAM解决方案。我们发布源代码，不仅是为了SLAM社区的利益，也是为了为其他领域的研究人员提供开箱即用的SLAM解决方案。
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.