Robust odometry estimation for RGB-D cameras

By Christian Kerl, Jürgen Sturm, Daniel Cremers. 2013

610 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): “Six research publications about evaluating ego-motion estimation and SLAM over TUM Benchmark dataset are [21, 38, 45, 84, 86, 87].”
A Large Dataset of Object Scans (2016, background): “[3], which aims to minimize the photometric error between consecutive RGB-D frames.”
A benchmark for RGB-D visual odometry, 3D reconstruction and SLAM (2014, methodology): “We evaluate 5 different odometers; (i) DVO [13] which performs full frame alignment between pairs of RGB-D frames making use of a robust cost function to improve robustness to outliers; (ii) FOVIS [14] which uses FAST feature correspondences between pairs of RGB-D frames to resolve camera motion, re”
Deeper Depth Prediction with Fully Convolutional Residual Networks (2016, methodology): “In particular, we deploy a SLAM framework where frame-to-frame tracking is obtained via GaussNewton optimization on the pixelwise intensity differences computed on consecutive frame pairs as proposed in [12], while fusion of depth measurements between the current frame and the global model is carrie”
PROBE-GK: Predictive robust estimation using generalized kernels (2016, methodology): “M-estimation) based on Student’s t weighting, with ν = 5 (as suggested in [2]).”
Improving RGB-D SLAM in dynamic environments: A motion removal approach (2017, background): “prone to be influenced by large or occasional errors [43].”
Dense visual SLAM for RGB-D cameras (2013, methodology): “In our previous work [7] we gave a probabilistic formulation for the estimation of the camera motion ξ given the photometric error rI .”
Direct visual-inertial odometry with stereo cameras (2016, methodology): “Since then, direct methods have been omni-present in the domain of RGB-D cameras [18] [27], as they directly provide the required pixel-wise depth as sensor measurement.”
Colored Point Cloud Registration Revisited (2017, methodology): “A photometric objective EI is formulated in terms of squared differences of intensities [35, 25]:”
SVO: Fast semi-direct monocular visual odometry (2014, methodology): “In practice, the distribution has heavier tails due to occlusions and thus, a robust cost function must be applied [10].”
Visual-lidar odometry and mapping: low-drift, robust, and fast (2015, methodology): “The visual odometry method used in our system is similar to [8]– [12] in the sense that all use visual images with additionally provided depth.”
Large-Scale Multi-resolution Surface Reconstruction from RGB-D Sequences (2013, methodology): “Non-linear Minimization We adapt the probabilistic framework proposed in [10] to combine the photometric and geometric error.”
Stereo DSO: Large-Scale Direct Sparse Visual Odometry with Stereo Cameras (2017, methodology): “Algorithmically they typically rely on robust cost functions and optimization through Gauss-Newton iteration as done also for RGBD based SLAM in [14, 15, 18].”
Robust visual inertial odometry using a direct EKF-based approach (2015, methodology): “By employing highly optimized SSE-vectorized implementations, ﬁrst real-time, CPU-based approaches for dense-or semi-dense motion estimation using a RGBD [15] or a monocular RGB camera [6], [7] have recently been proposed.”
PL-SVO: Semi-direct Monocular Visual Odometry by combining points and line segments (2016, methodology): “Solutions for the VO problem has been addressed employing diﬀerent sensors, such as monocular or stereo cameras [2] [3] [4], RGB-D cameras [5] [6], or a combination of any of them with an Inertial Measurement Unit (IMU) [7].”
Visual SLAM algorithms: a survey from 2010 to 2016 (2017, methodology): “On the other hand, phtometric consistency based camera motion tracking methods have been proposed [22, 23, 64].”
Robust reconstruction of indoor scenes (2015, methodology): “To derive more reliable information on local surface geometry, we partition the input RGB-D video into k-frame segments (k=50 in all experiments), use RGB-D odometry to estimate the camera trajectory [35], and fuse the range images to obtain a surface mesh for each segment [14].”
Joint SFM and detection cues for monocular 3D localization in road scenes (2015, methodology): “An important aspect of our object localization is the use of 3D points as SFM cues, for which we use dense tracking that exploits intensity-aligned pose optimization [9].”
SplaTAM: Splat, Track & Map 3D Gaussians for Dense RGB-D SLAM (2023, background): “Traditional approaches to dense SLAM have explored a variety of explicit representations, including 2.5D images [15, 34], (truncated) signed-distance functions [6, 25, 26, 42], gaussian mixture models [9, 10] and circular sur-fels [13, 33, 43].”