ORB-SLAM: A Versatile and Accurate Monocular SLAM System
By Raul Mur-Artal, José M. M. Montiel, Juan D. Tardós. 2015
7,382 citations recorded by the source corpus.
This paper presents ORB-SLAM, a feature-based monocular simultaneous localization and mapping (SLAM) system that operates in real time, in small and large indoor and outdoor environments. The system is robust to severe motion clutter, allows wide baseline loop closing and relocalization, and includes full automatic initialization. Building on excellent algorithms of recent years, we designed from scratch a novel system that uses the same features for all SLAM tasks: tracking, mapping, relocalization, and loop closing. A survival of the fittest strategy that selects the points and keyframes of the reconstruction leads to excellent robustness and generates a compact and trackable map that only grows if the scene content changes, allowing lifelong operation. We present an exhaustive evaluation in 27 sequences from the most popular datasets. ORB-SLAM achieves unprecedented performance with respect to other state-of-the-art monocular SLAM approaches. For the benefit of the community, we make the source code public.
Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition, Ocean Engineering
How researchers see this paper
The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.
In Aerospace Engineering
- Marc Pollefeys
- Andrew J. Davison
- Vladlen Koltun
- Daniel Cremers
- Davide Scaramuzza
- Torsten Sattler
- Andreas Geiger
- Jia Deng
In Computer Vision and Pattern Recognition
- Luc Van Gool
- Dacheng Tao
- Jitendra Malik
- Ming-Hsuan Yang
- Trevor Darrell
- Xiaogang Wang
- Andrea Vedaldi
- Andrew Zisserman
In Ocean Engineering
- Andrew J. Davison
- Daniel Cremers
- Juan D. Tardós
- John J. Leonard
- Davide Scaramuzza
- Shaojie Shen
- Michael Milford
- Andrea Vedaldi
Citations in the corpus
- A Survey of Deep Network Solutions for Learning Control in Robotics: From Reinforcement to Imitation (2016, neutral): “ the raw sensor readings based on artificially designed complex features (e.g., SIFT (Lowe, 1999, 2004; Liu et al., 2014), SURF (Bay et al., 2006, 2008), ORB (Rublee et al., 2011; Moreno et al., 2015; Mur-Artal et al., 2015)). Most of the classical methods are constrained by the adaptivity to generic”
- Deep Drone Racing: Learning Agile Flight in Dynamic Environments (2018, background): “Additionally, state-of-the-art state estimation pipelines may require expensive sensors [2], have high computational costs [22], or be subject to drift due to the use of compressed maps [18].”
- IQA: Visual Question Answering in Interactive Environments (2017, background): “Online map-based methods [10, 13, 50, 61, 67, 69] often construct the map while exploring the environment.”
- Unsupervised Learning of Depth and Ego-Motion from Video (2017, methodology): “Absolute Trajectory Error (ATE) on the KITTI odometry split averaged over all 5 -frame snippets (lower is better).”
- Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age (2016, positive): “They allow to build accurate and robust SLAM systems with automatic relocation and loop closing [179].”
- Playing for Benchmarks (2017, background): “We benchmark two state-of-the-art systems that represent different approaches to monocular visual odometry: ORB-SLAM2, which tracks sparse features [45], and DSO, which optimizes a photometric objective defined over image intensities [15].”
- The Replica Dataset: A Digital Replica of Indoor Spaces (2019, methodology): “The wide-angle greyscale video data together with the IMU data is used by an in-house SLAM system, similar to state-of-the-art systems like [10], [22], to provide 6 degree of freedom (DoF) poses.”
- AutoScaler: Scale-Attention Networks for Visual Correspondence (2016, background): “It is crucial for a plethora of applications, including optical flow [54, 43, 36], structure-from-motion [1], visual SLAM [40, 27, 39], stereo matching [59, 34], non-rigid 3D reconstruction [14] as well as video segmentation [19].”
- Neural Topological SLAM for Visual Navigation (2020, background): “Researchers have also tackled the problem of passive and active localization [7, 26], in order to aid building such consistent metric representations.”
- Large-scale direct SLAM for omnidirectional cameras (2015, methodology): “For comparison we also run the recently presented ORBSLAM [24] on the rectified dataset (Orb).”
- Ego-Topo: Environment Affordances From Egocentric Video (2020, methodology): “(x, y) locations are obtained from monocular SLAM [50], and the cell size is based on the typical scale of an interaction area following prior work [20].”
- Neural Autonomous Navigation with Riemannian Motion Policy (2019, background): “Although significant progress has been made [8], [9], [10], [11], these techniques still have trouble with textureless environments or fast-moving cameras.”
- Sparse-to-Dense: Depth Prediction from Sparse Depth Samples and a Single Image (2017, methodology): “4A typical feature-based SLAM algorithm, such as ORB-SLAM [7], keeps track of hundreds of 3D landmarks in each frame.”
- Geometric Consistency for Self-Supervised End-to-End Visual Odometry (2018, methodology): “Moreover, CTCNet alleviates the need for hyperparameter tuning, which is frequently required for traditional VO pipelines such as ORB-SLAM [17].”
- Kimera: an Open-Source Library for Real-Time Metric-Semantic Localization and Mapping (2019, background): “First, while Back-end Geometry Sema-ntics ORB-SLAM [22] mono g2o points DSO [23] mono g2o points VINS-mono [24] mono/IMU Ceres points VINS-Fusion [25] mono/Stereo/IMU Ceres points ROVIOLI [26] stereo/IMU EKF points ElasticFusion [18] RGB-D alternation surfels Voxblox [27] RGB-D [26] TSDF SLAM++…”
- Aggressive Deep Driving: Model Predictive Control with a CNN Cost Model (2017, methodology): “There are many SLAM approaches that use cameras [8, 9], LIDAR [10], or other sensor combinations [11] to provide accurate position.”
- Geometry-Aware Learning of Maps for Camera Localization (2017, methodology): “In visual SLAM, 3D landmarks (with feature descriptors of the corresponding image patches) are often defined as maps in both Bayesian filtering approaches [14, 41, 53] and key-frame based approaches [34, 42, 49].”
- Efficient Global 2D-3D Matching for Camera Localization in a Large-Scale 3D Map (2017, background): “Getting accurate estimation of 6-DoF camera pose from an image is essential for many computer vision applications such as robot navigation [15, 37], augmented reality [35, 55], and image-based 3D reconstruction [4,18].”