Software de procesamiento automático de placas espectrográficas

PEREYRA, N.; PONTE AHON, S.; AIDELMAN, Y.J.; RONCHETTI, F.; QUIROGA, F.; GAMEN, R; CIDALE, L.

CABA

Congreso; Reunión Anual de la AAA; 2022

AAA

The Faculty of Astronomical and Geophysical Sciences, through the Recovery of Historical Observational Work (ReTrOH, by its acronym in Spanish) project, is in the process of digitizing a large collection of glass-format spectroscopic plates. On the other hand, Neural Networks are currently the best performing Machine Learning models capable of solving a wide variety of problems. They are general models and universal approximators. In recent years, multi-layered Neural Networks have been successfully trained using a set of techniques often referred to as Deep Learning. In this context, we are developing a software for automatic processing of spectrographic plates, which detects the individual science spectra on these plates using Deep Learning and allows, in addition, to load their respective metadata.