DocumentCode
2714793
Title
Visibility Based Preconditioning for bundle adjustment
Author
Kushal, Avanish ; Agarwal, Sameer
Author_Institution
Univ. of Washington, Seattle, WA, USA
fYear
2012
fDate
16-21 June 2012
Firstpage
1442
Lastpage
1449
Abstract
We present Visibility Based Preconditioning (VBP) a new technique for efficiently solving the linear least squares problems that arise in bundle adjustment (Triggs et al., 1999). Using the camera-point visibility structure of the scene, we describe the construction of two preconditioners. These preconditioners when combined with an inexact step Levenberg-Marquardt algorithm (Wright and Holt, 1985) offer state of the art performance on the BAL data set (Agarwal et al., 2010), with 3-5× reduction in execution time over currently available methods while delivering comparable or better solution quality.
Keywords
cameras; image matching; image reconstruction; BAL data set; bundle adjustment; inexact step Levenberg-Marquardt algorithm; linear least squares problems; scene camera-point visibility structure; visibility based preconditioning; Approximation methods; Cameras; Clustering algorithms; Jacobian matrices; Symmetric matrices; Vectors; Vegetation;
fLanguage
English
Publisher
ieee
Conference_Titel
Computer Vision and Pattern Recognition (CVPR), 2012 IEEE Conference on
Conference_Location
Providence, RI
ISSN
1063-6919
Print_ISBN
978-1-4673-1226-4
Electronic_ISBN
1063-6919
Type
conf
DOI
10.1109/CVPR.2012.6247832
Filename
6247832
Link To Document