Back to Volume
Paper: Automated Image Registration for the Future
Volume: 394, Astronomical Data Analysis Software and Systems (ADASS) XVII
Page: 35
Authors: Hack, W.J.
Abstract: A primary problem facing all surveys or archives of astronomical images remains the automated registration of the images. Images often have pointing errors not accounted for in their headers or metadata, errors which need to be removed in order to successfully combine them into a single, deeper, more scientifically valuable product. The primary techniques rely on either matching catalogs of positions or performing cross-correlation on images. Each of these techniques can only be applied successfully to limited sets of astronomical data. This talk will review the primary techniques currently used for image registration and identify their most obvious limitations for use in automated registration. A new algorithm merging the best of these techniques with a proven technique developed outside of astronomy will be explored as an example of a new paradigm for solving the problem of automated and robust image registration.
Back to Volume