Automated detection of very low surface brightness galaxies in the Virgo cluster

We report the automatic detection of a new sample of very low surface brightness (LSB) galaxies, likely members of the Virgo cluster. We introduce our new software, DeepScan, that has been designed specifically to detect extended LSB features automatically using the DBSCAN algorithm. We demonstrate the technique by applying it over a 5 deg(2) portion of the Next Generation Virgo Survey (NGVS) data to reveal 53 LSB galaxies that are candidate cluster members based on their sizes and colours. 30 of these sources are new detections despite the region being searched specifically for LSB galaxies previously. Our final sample contains galaxies with 26.0 < (mu(e)) <= 28.5 and 19 < m(g) <= 21, making them some of the faintest known in Virgo. The majority of them have colours consistent with the red sequence, and have a mean stellar mass of 10(6.3 +/- 10.5) M-circle dot assuming cluster membership. After using ProFit to fit Sersic profiles to our detections, none of the new sources have effective radii larger than 1.5 Kpc and do not meet the criteria for ultra-diffuse galaxy (UDG) classification, so we classify them as ultra-faint dwarfs.