An implementation of the partitioned Levenberg-Marquardt algorithm for applications in computer vision.
DOI:
https://doi.org/10.5433/1679-0375.2009v30n1p51Keywords:
Levenberg-Marquardt algorithm, Monocular Calibration, Newton's.Abstract
At several applications of computer vision is necessary to estimate parameters for a specific model which best fits an experimental data set. For these cases, a minimization algorithm might be used and one of the most popular is the Levenberg-Marquardt algorithm. Although several free applies from this algorithm are available, any of them has great features when the resolution of problem has a sparse Jacobian matrix . In this case, it is possible to have a great reduce in the algorithm's complexity. This work presents a Levenberg-Marquardt algorithm implemented in cases which has a sparse Jacobian matrix. To illustrate this algorithm application, the camera calibration with 1D pattern is applied to solve the problem. Empirical results show that this method is able to figure out satisfactorily with few iterations, even with noise presence.
Downloads
Downloads
Published
How to Cite
Issue
Section
License
The Copyright Declaration for articles published in this journal is the author’s right. Since manuscripts are published in an open access Journal, they are free to use, with their own attributions, in educational and non-commercial applications. The Journal has the right to make, in the original document, changes regarding linguistic norms, orthography, and grammar, with the purpose of ensuring the standard norms of the language and the credibility of the Journal. It will, however, respect the writing style of the authors. When necessary, conceptual changes, corrections, or suggestions will be forwarded to the authors. In such cases, the manuscript shall be subjected to a new evaluation after revision. Responsibility for the opinions expressed in the manuscripts lies entirely with the authors.
This journal is licensed with a license Creative Commons Attribution-NonCommercial 4.0 International.