Implementation of Artificial Intelligence methods to improve galaxies cataloging

Supervisor: Dr.L. Sarro

The traditional classification and cataloging of astronomical objects required some heavy manual steps based on the astronomer perception and classification criteria to determine the most probable separation of overlapped objects, to perform probabilistic cross-matching by identify the same object in the different bands, etc. The considerable increase of data rate in the late years is challenging this traditional method against a new approach strongly based on data pipeline processing architecture involving intense and intelligent computational capabilities.Therefore, the ultimate goal of this research work is to contribute to the increasing understanding of the Universe by expanding the knowledge acquisition capabilities in the area of astronomical data analysis with the use of artificial intelligence methodologies, mainly in the areas of data mining, rule-based classification system and artificial vision, being all these aspects integrated in a coordinated knowledge engineering framework. The main results and conclusions derived from each of the main blocks of the pipeline developed will be presented.
The foundations of the contributions presented as part of this research can be expanded to similar scientific scenarios in Astronomy or in other disciplines.

Maria José Márquez