9-13 June 2013
Koper, Slovenia
UTC timezone
Home > Timetable > Contribution details
PDF | XML

A NON-LOCAL VARIATIONAL APPROACH FOR P+XS IMAGE FUSION

Presented by Ms. Catalina SBERT
Type: Oral presentation
Track: Mathematical Methods in Image Processing

Content

In this work, we present a nonlocal variational model for fusing the information given by a panchromatic image at high resolution and the spectral multichannel at lower resolution in order to get the high resolution of the multispectral images. Our approach is based on the Caselles et al work \cite{bciv-2006}. In this paper, the authors give a functional model based on the main assumption that the geometry of the spectral images is contained in the topographic map of the panchromatic image. In our study, by the contrary, we assume that we can infer the geometry of the spectral channels at high resolution by using the NL-means operator, proposed initially for denoising purposes in \cite{bcm-2005}, to transport the high frequency information from the true panchromatic image to the high resolution of the multispectral images. The other two terms of the proposed model, first the relation between the panchromatic image to the spectral channels and, on the other, the relation between the low and high resolution of the spectral images, are the same as those proposed in \cite{bciv-2006}. We study theoretically the proposed functional model to get existence and uniqueness of the minimum. Finally, we show some experiments and we discuss the results. \begin{thebibliography}{99} \bibitem{bciv-2006} C. Ballester, V. Caselles, L. Igual, J. Verdera and B. Rouge. \textit{A variational model for P+XS image fusion}, International Journal of Computer Vision \textbf{69}(1), 43--58, 2006. \bibitem{bcm-2005} A Buades, B Coll and JM Morel. \textit{A review of image denoising algorithms, with a new one}, Multiscale Model. Simul., \textbf{4}(2), 490--530, 2005. \end{thebibliography}

Place

Location: Koper, Slovenia

Primary authors

More

Co-authors

More