An evaluation of the RGB-D SLAM system

By Felix Endres, Jürgen Hess, Nikolas Engelhard, Jürgen Sturm, Daniel Cremers, Wolfram Burgard. 2012

839 citations recorded by the source corpus.

Subfields: Aerospace Engineering, 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 Vision and Pattern Recognition

Citations in the corpus

RGB-D datasets using microsoft kinect or similar sensors: a survey (2016, background): “19 Sample images from TUM Benchmark dataset [21]”
SUN3D: A Database of Big Spaces Reconstructed Using SfM and Object Labels (2013, methodology): “Our system then uses these object annotations as constraints in a generalized bundle adjustment procedure (Section 5) to refine the camera pose.”
Real-time grasp detection using convolutional neural networks (2014, background): “Robotic systems increasingly leverage RGB-D sensors and data for tasks like object recognition [8], detection [9] [10], and mapping [11] [12].”
Deformable Shape Completion with Graph Convolutional Autoencoders (2017, background): “For static scenes, this problem was explored extensively, e.g., in [37, 38, 15].”
Visual Place Recognition: A Survey (2016, background): “While many systems use data from additional sensors such as lasers [98] or RGB-D cameras [126]–[128], geometric data can also be extracted from conventional cameras to allow metric calculation of the robot pose.”
Large-Scale Multi-resolution Surface Reconstruction from RGB-D Sequences (2013, background): “While impressive progress in this domain has been achieved over the last decade [2, 1, 9, 5], many of these approaches are based on visual keypoints that are reconstructed in 3D, which typically leads to sparse reconstructions in form of 3D point clouds.”
Robust odometry estimation for RGB-D cameras (2013, methodology): “Odometry methods based on visual features such as SIFT or SURF are too slow for being used in low-latency applications [6], [7].”
Unsupervised Intrinsic Calibration of Depth Sensors via SLAM (2013, methodology): “SLAM - Our SLAM approach is similar to that of RGBDSLAM [6, 5] or RGB-D Mapping [9, 8], but with some modifications.”
Visual SLAM algorithms: a survey from 2010 to 2016 (2017, background): “In the literature [62, 63], relative camera motion is estimated by tracking feature points between successive frames.”
Computer Vision and Image Understanding Computer Vision for Assistive Technologies (neutral): “These technological advancements have led many researchers to investigate the possibility to create a novel approach to SLAM that combines the scale information of 3D depth sensing with the strengths of visual features to create dense 3D environment representations ( Endres et al., 2012 ).”
SVO: Semidirect Visual Odometry for Monocular and Multicamera Systems (2017, positive): “Results for [16], [34], [41], and [68] were obtained from [16] and for PTAM”
Elastic Fragments for Dense Scene Reconstruction (2013, methodology): “While prior work relied on manual initial alignment [1], we found that an off-the-shelf SLAM system [5] was sufficient for our purposes.”
LSD-SLAM: Large-Scale Direct Monocular SLAM (2014, background): “Note that [14] and [7] use depth information from the sensor, while the others do not.”
Dense 3D semantic mapping of indoor scenes from RGB-D images (2014, background): “Note that this procedure can easily be replaced by a SLAM approach with loop closure (such as [23]) for improved results and is not the focus of our approach.”
Visual Language Maps for Robot Navigation (2022, methodology): “Classic methods for robot navigation [4], [5] build geometric maps for path planning and can parse goals from natural language commands [6], [7], but struggle to generalize to unseen instructions.”
A benchmark for the evaluation of RGB-D SLAM systems (2012, methodology): “7b, where both the RPE and the ATE were computed on various estimated trajectories from the RGB-D SLAM system.”
Volumetric 3D mapping in real-time on a CPU (2014, background): “I. INTRODUCTION Reconstructing the geometry and texture of the world in real-time from a sequence of images remains one of the keychallenges in computer vision and robotics.”