Abstract
ABSTRACT We have developed a procedure for the automatic selection of emission-line galaxies candidates from the digitization of objective prism plates. This procedure has been applied over two pair of direct and prism plates of the Universidad Complutense de Madrid (UCM) survey, digitized by the fast and high performance microdensitometer MAMA (Machine Automatique a Mesurer pour l'Astronomie) located at the Observatoire de Paris. The plates where obtained with the Schmidt Telescope at Calar Alto Observatory in Almeria (Spain). The instrumental setup used for the acquisition of the prism plates (IIIa-F emulsion and RG630 filter) allows to register the H-alpha +[NII] emission for galaxies up to z ≤ 0.045. Therefore the criteria developed for the automatic selection of ELGs candidates consist in identifying the H-alpha emission feature in the extracted monodimensional prism spectra. We have noted that there is not a nitid edge to distinguish between the spectra showing the H-alpha line in emission and the rest. Consequently we have studied the residual obtained after the estimation of the continuum and the variation of the slopes calculated over each spectrum in order to identify the line feature. The efficiency of this technique has been tested by comparing the results obtained applying the automatic procedure with those achieved after a careful visual search for the candidates, as performed up to date in the UCM and similar surveys. The results show us that the automatic procedure only ignore very extended and near saturated known galaxies with well resolved emitting regions located at the external regions. Spectroscopic observations of moderate dispersion for the whole set of candidates selected both automatic and visually show that the former has been able, not only to select the 100% of candidates with confirmed emission recognized also by visual means, but also, and what is more important, to identify a 29% more of confirmed ELGs, without lost of efficiency by the increase of spurious identifications. The digitization of the plates has allowed us to measure a set of accurate observational parameters as positions, magnitudes, sizes and redshifts that will permit to perform statistical analysis of this kind of objects before carrying out specific observations with larger telescopes. Since our plates have not sensitometric spots to calculate its characteristic curve, it has been necessary to obtain external calibrations by comparing with calibrated data offered by several catalogs. We have tested that the precision achieved using this method is mainly limited by the accuracy of the different catalog employed. The redshift of the candidates can be derived from the location of the H-alpha line on the prism spectra. Since they have no spectral features to refer the position of the Hα line, the positions of the objects in the direct plate and their transformations to the prism one have been used to provide a reference point which is dependent neither on magnitude, nor on color, nor spatial light distribution. A precision of 0.003 is reached when measuring redshifts by this method. Finally, the comparison of the candidate samples obtained from two plates of our survey covering the same area on the sky allows the study of the selection effects that affects the identification of the H-alpha line in photographic prism plates. The parameter EW X F of the emission H-alpha +[NII] can be used as a threshold that inform us about the unambiguous identification of the emission in the prism spectra, being 10-13 erg s-1 cm-2A the mean value of this parameter for the whole sample of the UCM survey. We have also developed an algorithm to simulate the expected objective-prism spectra for an object by using different instrumental setups. The simulation permits to investigate the different combinations of EW, fluxes and magnitude that allow the detection of the Hα line in the plates and will be used in the near future to predict the expected spectra by the substitution of the photographic emulsion with CCD detectors.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have