Ionized nebulae are essential tools to study star-forming galaxies. They can be used to trace properties, like, among others, star-formation rate, gas density, temperature and chemical abundances. Most of the nebulae belong to three classes: H II regions, planetary nebulae, and supernova remnants. They are the product of different physical processes, and they trace different properties of their host galaxy, even though, at a first glance, they all look similar. For this reason, several methods have been proposed in the literature to properly classify ionized them. However, most of these methods are strongly dataset dependent, and while qualitatively in agreement with each other, there is much variety regarding the quantitative constraints driving the classification.
The advent of new high spatial resolution integral field spectroscopy instruments, such as VLT/MUSE, offers the possibility to develop new, more detailed, and comprehensive classification criteria for ionized nebulae. In particular, the PHANGS-MUSE survey observed a large fraction of the star-forming disk of 19 of the closest star-forming galaxies observable by the southern hemisphere. This survey is expected to detect tens of thousands of nebulae, all with complete spectral and morphological information.
We are currently developing a new algorithm based on model comparison to classify these nebulae taking advantage of all the information recovered by MUSE. In this talk I will present the first result from our work and discuss how it will be possible to use the catalog of classified sources to define new, flexible criteria, that can be applied to different datasets.