Is astronomy, an important task is the identification of specific celestial objects. This task is often performed manually, by analyzing spectral data of the objects. Due to the advances in technology employed to observe these objects, specially of telescopes that are capable of capturing images of hundreds of millions of objects, the use of efficient computational methods to help in this identification process is necessary. This research project is within the context of an important international project, the J-PAS, planned to survey a large sky area by capturing multispectral images in the visible spectra. We aim to investigate and develop a method to separate stars and quasars from these multispectral images, by analysing spectral dimension only. For that, signal feature extraction, dimensionality reduction and supervised classification methods will be employed. It is expected that besides the methods per se, this project will also contribute to establish similar initiatives involving effective collaboration among researchers from the fields of Statistics, Computation and Astronomy.
News published in Agência FAPESP Newsletter about the scholarship: