Minimal Solvers for Rectifying from Radially-Distorted Scales and Change of Scales

Pritts, James; Kúkelová, Zuzana; Larsson, Viktor; Lochman, Yaroslava; Chum, Ondřej

doi:10.1007/s11263-019-01216-x

Cited by 11 publications

(7 citation statements)

References 51 publications

(268 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Pritts et al [19,20,22,21] proposed a suite of solvers that can jointly undistort and affinely-rectify from the imaged translation directions of coplanar repeated scene texture. The solvers directly estimate the vanishing line of the scene plane but also return the vanishing directions of the imaged translations that are consistent with the recovered vanishing line.…”

Section: Related Workmentioning

confidence: 99%

“…The corresponding points of two coplanar translated repeated regions [25] form parallel scene lines. Thus we can use the point correspondences extracted from distorted images of coplanar repeated texture [19,22] to construct the undistorted images of parallel lines. Let x and x be two distorted points in correspondence.…”

Section: Imaged Translated Coplanar Repeatsmentioning

confidence: 99%

“…Point correspondences are extracted from a covariant region correspondence [17,22]. In particular, we use the…”

Section: Featuresmentioning

confidence: 99%

“…In the presence of imaging noise, good feature coverage over large parts of the image is necessary to observe the joint effects of perspective warp and lens distortion. State-of-the-art techniques for auto-calibrating or rectifying lens-distorted images use either circular arcs fitted to edge detections or covariant region detections as inputs [1,33,22,21]. Complementary features can provide measurements on image regions that lack either tex-ture or lines.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Minimal Solvers for Single-View Lens-Distorted Camera Auto-Calibration

Lochman¹,

Dobosevych²,

Hryniv³

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

This paper proposes minimal solvers that use combinations of imaged translational symmetries and parallel scene lines to jointly estimate lens undistortion with either affine rectification or focal length and absolute orientation. We use constraints provided by orthogonal scene planes to recover the focal length. We show that solvers using feature combinations can recover more accurate calibrations than solvers using only one feature type on scenes that have a balance of lines and texture. We also show that the proposed solvers are complementary and can be used together in a RANSAC-based estimator to improve auto-calibration accuracy. State-of-the-art performance is demonstrated on a standard dataset of lens-distorted urban images. The code is available at https://github. com/ylochman/single-view-autocalib.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Imaged Translated Coplanar Repeatsmentioning

confidence: 99%

“…Point correspondences are extracted from a covariant region correspondence [17,22]. In particular, we use the…”

Section: Featuresmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Minimal Solvers for Single-View Lens-Distorted Camera Auto-Calibration

Lochman¹,

Dobosevych²,

Hryniv³

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…mal images, image rectification is often used as the preprocessing step to correct the distortions. Various traditional geometry-based methods have been proposed in the past decades which formulate image rectification as an optimization problem [5,1,41,27]. Recently, deep learningbased methods have shown promising results by leveraging the representation power of deep neural networks [29,40,20,21,38,39].…”

Section: Introductionmentioning

confidence: 99%

SIR: Self-supervised Image Rectification via Seeing the Same Scene from Multiple Different Lenses

Fan¹,

Zhang²,

Tao³

2020

Preprint

View full text Add to dashboard Cite

Deep learning has demonstrated its power in image rectification by leveraging the representation capacity of deep neural networks via supervised training based on a largescale synthetic dataset. However, the model may overfit the synthetic images and generalize not well on realworld fisheye images due to the limited universality of a specific distortion model and the lack of explicitly modeling the distortion and rectification process. In this paper, we propose a novel self-supervised image rectification (SIR) method based on an important insight that the rectified results of distorted images of the same scene from different lens should be the same. Specifically, we devise a new network architecture with a shared encoder and several prediction heads, each of which predicts the distortion parameter of a specific distortion model. We further leverage a differentiable warping module to generate the rectified images and re-distorted images from the distortion parameters and exploit the intra-and inter-model consistency between them during training, thereby leading to a self-supervised learning scheme without the need for ground-truth distortion parameters or normal images. Experiments on synthetic dataset and real-world fisheye images demonstrate that our method achieves comparable or even better performance than the supervised baseline method and representative state-of-the-art methods. Self-supervised learning also improves the universality of distortion models while keeping their self-consistency.

show abstract

Relative Pose from Deep Learned Depth and a Single Affine Correspondence

Eichhardt

Baráth

2020

Lecture Notes in Computer Science

View full text Add to dashboard Cite

We propose a new approach for combining deep-learned nonmetric monocular depth with affine correspondences (ACs) to estimate the relative pose of two calibrated cameras from a single correspondence. Considering the depth information and affine features, two new constraints on the camera pose are derived. The proposed solver is usable within 1-point RANSAC approaches. Thus, the processing time of the robust estimation is linear in the number of correspondences and, therefore, orders of magnitude faster than by using traditional approaches. The proposed 1AC+D 4 solver is tested both on synthetic data and on 110 395 publicly available real image pairs where we used an off-the-shelf monocular depth network to provide up-to-scale depth per pixel. The proposed 1AC+D leads to similar accuracy as traditional approaches while being significantly faster. When solving large-scale problems, e.g. pose-graph initialization for Structure-from-Motion (SfM) pipelines, the overhead of obtaining ACs and monocular depth is negligible compared to the speed-up gained in the pairwise geometric verification, i.e., relative pose estimation. This is demonstrated on scenes from the 1DSfM dataset using a state-of-the-art global SfM algorithm.

show abstract

Minimal Solvers for Rectifying from Radially-Distorted Scales and Change of Scales

Cited by 11 publications

References 51 publications

Minimal Solvers for Single-View Lens-Distorted Camera Auto-Calibration

Minimal Solvers for Single-View Lens-Distorted Camera Auto-Calibration

SIR: Self-supervised Image Rectification via Seeing the Same Scene from Multiple Different Lenses

Relative Pose from Deep Learned Depth and a Single Affine Correspondence

Contact Info

Product

Resources

About