Name
Abstract | ||
|---|---|---|
To obtain a Euclidean reconstruction from images the cameras have to be calibrated. In recent years different approaches have been proposed to avoid explicit calibration. The problem with these methods is that several parameters have to be retrieved at once. Because of the non-linearity of the equations this is not an easy task and the methods often fail to converge. In the's paper a stratified approach is proposed which allows to first retrieve the affine calibration of the camera using the modulus constraint. Having the affine calibration it is easy to upgrade to Euclidean. The important advantage of this method is that only three parameters have to be evaluated at first. From a practical point of view, the major gain is that an affine reconstruction is obtained from arbitrary sequences of views, whereas so far affine reconstruction has been based on pairs of views with a pure translation in between. A short illustration of another application is also given. Once the affine calibration is known, the constraint can be used to retrieve the Euclidean calibration in the presence of a variable focal length |
Year | DOI | Venue |
|---|---|---|
1996 | 10.1109/ICPR.1996.546047 | ICPR '96 Proceedings of the 1996 International Conference on Pattern Recognition (ICPR '96) Volume I - Volume 7270 |
Keywords | Field | DocType |
Modulus Constraint,New Constraint | Affine transformation,Affine shape adaptation,Affine space,Computer science,Artificial intelligence,Euclidean geometry,Computer vision,Mathematical optimization,Affine combination,Affine coordinate system,Algorithm,Focal length,Calibration | Conference |
Volume | ISSN | ISBN |
1 | 1051-4651 | 0-8186-7282-X |
Citations | PageRank | References |
38 | 6.66 | 0 |
Authors | ||
3 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Marc Pollefeys | 1 | 7671 | 475.90 |
| Luc Van Gool | 2 | 27566 | 1819.51 |
| Oosterlinck, A. | 3 | 711 | 254.21 |