Minimal solutions for pose estimation of a multi-camera system

Gim Hee Lee, Bo Li, Marc Pollefeys, Friedrich Fraundorfer

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

Abstract

In this paper, we propose a novel formulation to solve the pose estimation problem of a calibrated multi-camera system. The non-central rays that pass through the 3D world points and multi-camera system are elegantly represented as Plücker lines. This allows us to solve for the depth of the points along the Plücker lines with a minimal set of 3-point correspondences. We show that the minimal solution for the depth of the points along the Plücker lines is an 8 degree polynomial that gives up to 8 real solutions. The coordinates of the 3D world points in the multi-camera frame are computed from the known depths. Consequently, the pose of the multi-camera system, i.e. the rigid transformation between the world and multi-camera frames can be obtained from absolute orientation. We also derive a closed-form minimal solution for the absolute orientation. This removes the need for the computationally expensive Singular Value Decompositions (SVD) during the evaluations of the possible solutions for the depths. We identify the correct solution and do robust estimation with RANSAC. Finally, the solution is further refined by including all the inlier correspondences in a non-linear refinement step. We verify our approach by showing comparisons with other existing approaches and results from large-scale real-world datasets.

Original languageEnglish
Title of host publicationSpringer Tracts in Advanced Robotics
PublisherSpringer Verlag
Pages521-538
Number of pages18
Volume114
ISBN (Print)9783319288703
DOIs
Publication statusPublished - 2016
Event16th International Symposium of Robotics Research, ISRR 2013 - Singapore, Singapore
Duration: 16 Dec 201319 Dec 2013

Publication series

NameSpringer Tracts in Advanced Robotics
Volume114
ISSN (Print)16107438
ISSN (Electronic)1610742X

Conference

Conference16th International Symposium of Robotics Research, ISRR 2013
CountrySingapore
CitySingapore
Period16/12/1319/12/13

Fingerprint

Cameras
Singular value decomposition
Polynomials

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Artificial Intelligence

Cite this

Hee Lee, G., Li, B., Pollefeys, M., & Fraundorfer, F. (2016). Minimal solutions for pose estimation of a multi-camera system. In Springer Tracts in Advanced Robotics (Vol. 114, pp. 521-538). (Springer Tracts in Advanced Robotics; Vol. 114). Springer Verlag. https://doi.org/10.1007/978-3-319-28872-7_30

Minimal solutions for pose estimation of a multi-camera system. / Hee Lee, Gim; Li, Bo; Pollefeys, Marc; Fraundorfer, Friedrich.

Springer Tracts in Advanced Robotics. Vol. 114 Springer Verlag, 2016. p. 521-538 (Springer Tracts in Advanced Robotics; Vol. 114).

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

Hee Lee, G, Li, B, Pollefeys, M & Fraundorfer, F 2016, Minimal solutions for pose estimation of a multi-camera system. in Springer Tracts in Advanced Robotics. vol. 114, Springer Tracts in Advanced Robotics, vol. 114, Springer Verlag, pp. 521-538, 16th International Symposium of Robotics Research, ISRR 2013, Singapore, Singapore, 16/12/13. https://doi.org/10.1007/978-3-319-28872-7_30
Hee Lee G, Li B, Pollefeys M, Fraundorfer F. Minimal solutions for pose estimation of a multi-camera system. In Springer Tracts in Advanced Robotics. Vol. 114. Springer Verlag. 2016. p. 521-538. (Springer Tracts in Advanced Robotics). https://doi.org/10.1007/978-3-319-28872-7_30
Hee Lee, Gim ; Li, Bo ; Pollefeys, Marc ; Fraundorfer, Friedrich. / Minimal solutions for pose estimation of a multi-camera system. Springer Tracts in Advanced Robotics. Vol. 114 Springer Verlag, 2016. pp. 521-538 (Springer Tracts in Advanced Robotics).
@inproceedings{69d50ec1a08845b5b1a2a6e2240a48e7,
title = "Minimal solutions for pose estimation of a multi-camera system",
abstract = "In this paper, we propose a novel formulation to solve the pose estimation problem of a calibrated multi-camera system. The non-central rays that pass through the 3D world points and multi-camera system are elegantly represented as Pl{\"u}cker lines. This allows us to solve for the depth of the points along the Pl{\"u}cker lines with a minimal set of 3-point correspondences. We show that the minimal solution for the depth of the points along the Pl{\"u}cker lines is an 8 degree polynomial that gives up to 8 real solutions. The coordinates of the 3D world points in the multi-camera frame are computed from the known depths. Consequently, the pose of the multi-camera system, i.e. the rigid transformation between the world and multi-camera frames can be obtained from absolute orientation. We also derive a closed-form minimal solution for the absolute orientation. This removes the need for the computationally expensive Singular Value Decompositions (SVD) during the evaluations of the possible solutions for the depths. We identify the correct solution and do robust estimation with RANSAC. Finally, the solution is further refined by including all the inlier correspondences in a non-linear refinement step. We verify our approach by showing comparisons with other existing approaches and results from large-scale real-world datasets.",
author = "{Hee Lee}, Gim and Bo Li and Marc Pollefeys and Friedrich Fraundorfer",
year = "2016",
doi = "10.1007/978-3-319-28872-7_30",
language = "English",
isbn = "9783319288703",
volume = "114",
series = "Springer Tracts in Advanced Robotics",
publisher = "Springer Verlag",
pages = "521--538",
booktitle = "Springer Tracts in Advanced Robotics",
address = "Germany",

}

TY - GEN

T1 - Minimal solutions for pose estimation of a multi-camera system

AU - Hee Lee, Gim

AU - Li, Bo

AU - Pollefeys, Marc

AU - Fraundorfer, Friedrich

PY - 2016

Y1 - 2016

N2 - In this paper, we propose a novel formulation to solve the pose estimation problem of a calibrated multi-camera system. The non-central rays that pass through the 3D world points and multi-camera system are elegantly represented as Plücker lines. This allows us to solve for the depth of the points along the Plücker lines with a minimal set of 3-point correspondences. We show that the minimal solution for the depth of the points along the Plücker lines is an 8 degree polynomial that gives up to 8 real solutions. The coordinates of the 3D world points in the multi-camera frame are computed from the known depths. Consequently, the pose of the multi-camera system, i.e. the rigid transformation between the world and multi-camera frames can be obtained from absolute orientation. We also derive a closed-form minimal solution for the absolute orientation. This removes the need for the computationally expensive Singular Value Decompositions (SVD) during the evaluations of the possible solutions for the depths. We identify the correct solution and do robust estimation with RANSAC. Finally, the solution is further refined by including all the inlier correspondences in a non-linear refinement step. We verify our approach by showing comparisons with other existing approaches and results from large-scale real-world datasets.

AB - In this paper, we propose a novel formulation to solve the pose estimation problem of a calibrated multi-camera system. The non-central rays that pass through the 3D world points and multi-camera system are elegantly represented as Plücker lines. This allows us to solve for the depth of the points along the Plücker lines with a minimal set of 3-point correspondences. We show that the minimal solution for the depth of the points along the Plücker lines is an 8 degree polynomial that gives up to 8 real solutions. The coordinates of the 3D world points in the multi-camera frame are computed from the known depths. Consequently, the pose of the multi-camera system, i.e. the rigid transformation between the world and multi-camera frames can be obtained from absolute orientation. We also derive a closed-form minimal solution for the absolute orientation. This removes the need for the computationally expensive Singular Value Decompositions (SVD) during the evaluations of the possible solutions for the depths. We identify the correct solution and do robust estimation with RANSAC. Finally, the solution is further refined by including all the inlier correspondences in a non-linear refinement step. We verify our approach by showing comparisons with other existing approaches and results from large-scale real-world datasets.

UR - http://www.scopus.com/inward/record.url?scp=84964916759&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-28872-7_30

DO - 10.1007/978-3-319-28872-7_30

M3 - Conference contribution

SN - 9783319288703

VL - 114

T3 - Springer Tracts in Advanced Robotics

SP - 521

EP - 538

BT - Springer Tracts in Advanced Robotics

PB - Springer Verlag

ER -