Share Email Print

Proceedings Paper

Deconvolution of astronomical data: What for and where do we go?
Author(s): Eric Pantin
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

Modern astronomy tries to push the observation instruments to their limits in order to discover new unsuspected phenomena in the universe. However, some intrinsic limitations (diffraction or atmospheric seeing etc ...) that degrade strongly the spatial resolution when imaging objects in the sky, and can be hardly reduced simply by technology improvements. In order to get the best scientific return from new, more and more sensitive instruments, image deconvolution methods try to push back these limits. Indeed, when the spatial degradation is known, partly known, or even unknown, deconvolution algorithms have proven to be able to restore an original image that is close, within the errors of the method, to a perfect image as observed by a perfect instrument with no, or heavily reduced image degradation. Many methods and algorithms exist nowadays to solve numerically the problem of image deconvolution. I will review the most popular ones and show their characteristics and limitations. In particular, I will show how we moved from standard methods (which see the images as one unique object) to multiscale methods that analyse the data from a multiresolution point a view, decoupling an originally very complicated problem into a set of problems more easy to solve.

Paper Details

Date Published: 19 December 2002
PDF: 12 pages
Proc. SPIE 4847, Astronomical Data Analysis II, (19 December 2002); doi: 10.1117/12.461973
Show Author Affiliations
Eric Pantin, CEA Saclay (France)

Published in SPIE Proceedings Vol. 4847:
Astronomical Data Analysis II
Jean-Luc Starck; Fionn D. Murtagh, Editor(s)

© SPIE. Terms of Use
Back to Top