Advanced

Conjugate Gradient Bundle Adjustment

Byröd, Martin LU and Åström, Kalle LU (2010) 11th European Conference on Computer Vision In Computer Vision-Eccv 2010, Pt II 6312. p.114-127
Abstract
Bundle adjustment for multi-view reconstruction is traditionally done using the Levenberg-Marquardt algorithm with a direct linear solver, which is computationally very expensive. An alternative to this approach is to apply the conjugate gradients algorithm in the inner loop. Tins is appealing since the main computational step of the CG algorithm involves only a simple matrix-vector multiplication with the Jacobian. In this work we improve on the latest published approaches to bundle adjustment with conjugate gradients by making full use of the least squares nature of the problem. We employ an easy-to-compute QR factorization based block preconditioner and show how a certain property of the preconditioned system allows us to reduce the... (More)
Bundle adjustment for multi-view reconstruction is traditionally done using the Levenberg-Marquardt algorithm with a direct linear solver, which is computationally very expensive. An alternative to this approach is to apply the conjugate gradients algorithm in the inner loop. Tins is appealing since the main computational step of the CG algorithm involves only a simple matrix-vector multiplication with the Jacobian. In this work we improve on the latest published approaches to bundle adjustment with conjugate gradients by making full use of the least squares nature of the problem. We employ an easy-to-compute QR factorization based block preconditioner and show how a certain property of the preconditioned system allows us to reduce the work per iteration to roughly half of the standard CG algorithm. (Less)
Please use this url to cite or link to this publication:
author
organization
publishing date
type
Chapter in Book/Report/Conference proceeding
publication status
published
subject
in
Computer Vision-Eccv 2010, Pt II
volume
6312
pages
114 - 127
publisher
Springer
conference name
11th European Conference on Computer Vision
external identifiers
  • wos:000286164000009
  • scopus:78149333899
ISSN
0302-9743
1611-3349
language
English
LU publication?
yes
id
2dce9d2b-d564-4f65-b419-0aa64b2d1c24 (old id 1859660)
date added to LUP
2011-04-20 10:50:05
date last changed
2018-07-01 03:13:12
@inproceedings{2dce9d2b-d564-4f65-b419-0aa64b2d1c24,
  abstract     = {Bundle adjustment for multi-view reconstruction is traditionally done using the Levenberg-Marquardt algorithm with a direct linear solver, which is computationally very expensive. An alternative to this approach is to apply the conjugate gradients algorithm in the inner loop. Tins is appealing since the main computational step of the CG algorithm involves only a simple matrix-vector multiplication with the Jacobian. In this work we improve on the latest published approaches to bundle adjustment with conjugate gradients by making full use of the least squares nature of the problem. We employ an easy-to-compute QR factorization based block preconditioner and show how a certain property of the preconditioned system allows us to reduce the work per iteration to roughly half of the standard CG algorithm.},
  author       = {Byröd, Martin and Åström, Kalle},
  booktitle    = {Computer Vision-Eccv 2010, Pt II},
  issn         = {0302-9743},
  language     = {eng},
  pages        = {114--127},
  publisher    = {Springer},
  title        = {Conjugate Gradient Bundle Adjustment},
  volume       = {6312},
  year         = {2010},
}