G2o: A general framework for graph optimization

By Rainer Kümmerle, Giorgio Grisetti, Hauke Strasdat, Kurt Konolige, Wolfram Burgard. 2011

2,451 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Networks and Communications, Computer Vision and Pattern Recognition

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

In Computer Networks and Communications

In Computer Vision and Pattern Recognition

Citations in the corpus

A Survey of Deep Network Solutions for Learning Control in Robotics: From Reinforcement to Imitation (2016, neutral): “Inspired by graph-based SLAM algorithms (Thrun et al., 2005; K¨ummerle et al., 2011”
The limits and potentials of deep learning for robotics (2018, methodology): “This would allow robots to treat a deep neural network like any other sensor, and use the established Bayesian techniques [53], [58], [117] to fuse the network’s predictions with prior knowledge or other sensor measurements, or to accumulate information over time.”
Past, Present, and Future of Simultaneous Localization and Mapping: Toward the Robust-Perception Age (2016, positive): “, GTSAM [62], g2o [146], Ceres [214], iSAM [126], and SLAM++ [204]) are able to solve problems with tens of thousands of variables in few seconds.”
A General Optimization-based Framework for Local Odometry Estimation with Multiple Sensors (2019, methodology): “Early optimization solvers, such as G2O [19], leveraged the Gauss-Newton and Levenberg-Marquardt approaches to solve the problem.”
Graduated Non-Convexity for Robust Spatial Perception: From Non-Minimal Solvers to Global Outlier Rejection (2019, methodology): “We benchmark GNC-GM and GNC-TLS against (i) g2o [20], (ii) dynamic covariance scaling (DCS) [42], (iii) pairwise consistent measurement set maximization (PCM) [30], and (iv)ADAPT [27].”
Multiple‐Robot Simultaneous Localization and Mapping: A Review (2016, methodology): “For instance, in (Henry, Krainin, Herbst, Ren, & Fox, 2010) and (Engelhard, Endres, Hess, Sturm, & Burgard, 2011), a method called rgbdSLAM is proposed to perform 3D SLAM using stereo and Kinect Journal of Field Robotics DOI 10.1002/rob cameras.”
On behalf of: Multimedia Archives (2012, positive): “Furthermore, a sparse LM method as implemented in the g 2 o library (Kümmerle et al., 2011) is employed.”
Unsupervised Intrinsic Calibration of Depth Sensors via SLAM (2013, background): “It is this latter case that we will explore in the following sections.”
Robust visual SLAM across seasons (2015, positive): “employ the go framework [14].”
Keyframe-Based Visual-Inertial SLAM using Nonlinear Optimization (2013, methodology): “We use g(2)o [11] as an optimization framework.”
CNN-SLAM: Real-Time Dense Monocular SLAM with Learned Depth Prediction (2017, background): “Moreover, the pose of the key-frames is each time globally refined via pose graph optimization [14].”
Automatic Registration of RGB-D Scans via Salient Directions (2013, methodology): “We execute the former but refer for details to [13], since the focus of this work is on the initial pair-wise registration.”
Real-Time Large-Scale Dense 3D Reconstruction with Loop Closure (2016, methodology): “Upon loop closure, the accumulated error is distributed by applying pose graph optimisation [30].”
Robust odometry estimation for RGB-D cameras (2013, background): “…on RGB-D data, • a robust sensor model derived from real world data, • the integration of a temporal prior, • an open-souce implementation that runs in real-time (30 Hz) on a single CPU core, • the rigorous evaluation on benchmark data demonstrating the high accuracy and robustness of our approach.”
MonoPair: Monocular 3D Object Detection Using Pairwise Spatial Relationships (2020, methodology): “We use g2o [18] to conduct this graph optimization structure during implementation.”
Double window optimisation for constant time visual SLAM (2011, methodology): “Both methods are implemented using the efficient state-of-the-art sparse graph optimizer g(2)o [9] and executed on a single core of an Intel(R) Core(TM) i7 960 desktop computer.”
A benchmark for the evaluation of RGB-D SLAM systems (2012, background): “from-motion) problem has a long history both in robotics [6]–[12] and in computer vision [9], [13]–[16].”
Global Pose Estimation with an Attention-Based Recurrent Network (2018, methodology): “The Neural Graph Optimizer model is a neural analogue of the global optimization procedures commonly used in traditional state-of-the-art SLAM packages, such the g2o framework [25].”