1995A&AS..113..399A

A full set of algorithms for the automatic analysis of low-resolution objective-prism spectra has been developed with the aim of finding new emission-line galaxies (ELGs) candidates from the Universidad Complutense de Madrid (UCM) survey prism plates. An objective-prism plate and a direct plate, used for pairing objects and to perform star-galaxy separation, have been scanned with the MAMA machine. The plates were taken in the red region of the spectrum, using the Hα+[NII] blend in emission as selection criterion. The procedure, applied to digitized spectra, is able to automatically select ELGs candidates. Density to intensity transformation has not been used in our method; the detection of the emission is performed applying three independent criteria over the one-dimensional spectra in raw data. The automatically selected sample is compared with that obtained after a careful visual scan. Spectroscopic observations at moderate spectral resolution are presented for the whole sample of candidates in order to study the drawbacks and the biases of both methods.