Simultaneous Multi-view Relative Pose Estimation and 3D Reconstruction from Planar Regions

Robert Frohlich, Z. Kato

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In this paper, we propose a novel solution for multi-view reconstruction, relative pose and homography estimation using planar regions. The proposed method doesn‘t require point matches, it directly uses a pair of planar image regions and simultaneously reconstructs the normal and distance of the corresponding 3D planar surface patch, the relative pose of the cameras as well as the aligning homography between the image regions. When more than two cameras are available, then a special region-based bundle adjustment is proposed, which provides robust estimates in a multi-view camera system by constructing and solving a non-linear system of equations. The method is quantitatively evaluated on a large synthetic dataset as well as on the KITTI vision benchmark dataset.

Original languageEnglish
Title of host publicationComputer Vision – ACCV 2018 Workshops - 14th Asian Conference on Computer Vision, 2018, Revised Selected Papers
EditorsGustavo Carneiro, Shaodi You
PublisherSpringer Verlag
Pages467-483
Number of pages17
ISBN (Print)9783030210731
DOIs
Publication statusPublished - Jan 1 2019
Event14th Asian Conference on Computer Vision, ACCV 2018 - Perth, Australia
Duration: Dec 2 2018Dec 6 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11367 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference14th Asian Conference on Computer Vision, ACCV 2018
CountryAustralia
CityPerth
Period12/2/1812/6/18

Fingerprint

Pose Estimation
3D Reconstruction
Cameras
Homography
Camera
Robust Estimate
Nonlinear Systems of Equations
Nonlinear systems
Patch
Bundle
Adjustment
Benchmark

Keywords

  • 3D reconstruction
  • Planar regions
  • Pose estimation

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Frohlich, R., & Kato, Z. (2019). Simultaneous Multi-view Relative Pose Estimation and 3D Reconstruction from Planar Regions. In G. Carneiro, & S. You (Eds.), Computer Vision – ACCV 2018 Workshops - 14th Asian Conference on Computer Vision, 2018, Revised Selected Papers (pp. 467-483). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11367 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-21074-8_37

Simultaneous Multi-view Relative Pose Estimation and 3D Reconstruction from Planar Regions. / Frohlich, Robert; Kato, Z.

Computer Vision – ACCV 2018 Workshops - 14th Asian Conference on Computer Vision, 2018, Revised Selected Papers. ed. / Gustavo Carneiro; Shaodi You. Springer Verlag, 2019. p. 467-483 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11367 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Frohlich, R & Kato, Z 2019, Simultaneous Multi-view Relative Pose Estimation and 3D Reconstruction from Planar Regions. in G Carneiro & S You (eds), Computer Vision – ACCV 2018 Workshops - 14th Asian Conference on Computer Vision, 2018, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11367 LNCS, Springer Verlag, pp. 467-483, 14th Asian Conference on Computer Vision, ACCV 2018, Perth, Australia, 12/2/18. https://doi.org/10.1007/978-3-030-21074-8_37
Frohlich R, Kato Z. Simultaneous Multi-view Relative Pose Estimation and 3D Reconstruction from Planar Regions. In Carneiro G, You S, editors, Computer Vision – ACCV 2018 Workshops - 14th Asian Conference on Computer Vision, 2018, Revised Selected Papers. Springer Verlag. 2019. p. 467-483. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-21074-8_37
Frohlich, Robert ; Kato, Z. / Simultaneous Multi-view Relative Pose Estimation and 3D Reconstruction from Planar Regions. Computer Vision – ACCV 2018 Workshops - 14th Asian Conference on Computer Vision, 2018, Revised Selected Papers. editor / Gustavo Carneiro ; Shaodi You. Springer Verlag, 2019. pp. 467-483 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{d9e4fb5c8b844c6c82ff2761950e1150,
title = "Simultaneous Multi-view Relative Pose Estimation and 3D Reconstruction from Planar Regions",
abstract = "In this paper, we propose a novel solution for multi-view reconstruction, relative pose and homography estimation using planar regions. The proposed method doesn‘t require point matches, it directly uses a pair of planar image regions and simultaneously reconstructs the normal and distance of the corresponding 3D planar surface patch, the relative pose of the cameras as well as the aligning homography between the image regions. When more than two cameras are available, then a special region-based bundle adjustment is proposed, which provides robust estimates in a multi-view camera system by constructing and solving a non-linear system of equations. The method is quantitatively evaluated on a large synthetic dataset as well as on the KITTI vision benchmark dataset.",
keywords = "3D reconstruction, Planar regions, Pose estimation",
author = "Robert Frohlich and Z. Kato",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/978-3-030-21074-8_37",
language = "English",
isbn = "9783030210731",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "467--483",
editor = "Gustavo Carneiro and Shaodi You",
booktitle = "Computer Vision – ACCV 2018 Workshops - 14th Asian Conference on Computer Vision, 2018, Revised Selected Papers",

}

TY - GEN

T1 - Simultaneous Multi-view Relative Pose Estimation and 3D Reconstruction from Planar Regions

AU - Frohlich, Robert

AU - Kato, Z.

PY - 2019/1/1

Y1 - 2019/1/1

N2 - In this paper, we propose a novel solution for multi-view reconstruction, relative pose and homography estimation using planar regions. The proposed method doesn‘t require point matches, it directly uses a pair of planar image regions and simultaneously reconstructs the normal and distance of the corresponding 3D planar surface patch, the relative pose of the cameras as well as the aligning homography between the image regions. When more than two cameras are available, then a special region-based bundle adjustment is proposed, which provides robust estimates in a multi-view camera system by constructing and solving a non-linear system of equations. The method is quantitatively evaluated on a large synthetic dataset as well as on the KITTI vision benchmark dataset.

AB - In this paper, we propose a novel solution for multi-view reconstruction, relative pose and homography estimation using planar regions. The proposed method doesn‘t require point matches, it directly uses a pair of planar image regions and simultaneously reconstructs the normal and distance of the corresponding 3D planar surface patch, the relative pose of the cameras as well as the aligning homography between the image regions. When more than two cameras are available, then a special region-based bundle adjustment is proposed, which provides robust estimates in a multi-view camera system by constructing and solving a non-linear system of equations. The method is quantitatively evaluated on a large synthetic dataset as well as on the KITTI vision benchmark dataset.

KW - 3D reconstruction

KW - Planar regions

KW - Pose estimation

UR - http://www.scopus.com/inward/record.url?scp=85068445341&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85068445341&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-21074-8_37

DO - 10.1007/978-3-030-21074-8_37

M3 - Conference contribution

AN - SCOPUS:85068445341

SN - 9783030210731

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 467

EP - 483

BT - Computer Vision – ACCV 2018 Workshops - 14th Asian Conference on Computer Vision, 2018, Revised Selected Papers

A2 - Carneiro, Gustavo

A2 - You, Shaodi

PB - Springer Verlag

ER -