|
Welcome to the OPERA Home Page
Download OPERA-1.3 now!
OPERA (Objective Prism Enhanced Reduction Algorithms) is a package of software tools for the authomatic analysis of astronomical images. It is being developed by the Universidad Complutense de Madrid (UCM, Spain) Astrophysics researh group. It is copyright by C.E. García-Dabó. Contact C.E. García-Dabó at ceg@astrax.fis.ucm.es for further information
Introduction
Surveys of star-forming galaxies are of vital importance to constrain
galactic evolution theories. One of the most successful searching
methods is the objective-prism (OP) technique, which can register
thousands of low resolution spectra in large areas. The UCM-CIDA
project expects to cover 150 square degrees and detect ~2000
H_alfa emission line galaxies.
OPERA is an automatic method to handle and analyze such amount of
data. It goes trough the detection of the objects in the image,
extraction of one-dimensional spectra and identification of the
emission line feature.
The main advantages of a method of this kind are: 1) to avoid
subjectivity inherent to visual inspection used in past studies; and
2) the ability to obtain physical parameters without follow-up
spectroscopy.
Description of the method
After common preprocessing (bias, flat, etc..) the first step is the
identification of sources on both direct and OP images. This
involves the previous determination of the local sky background and
associated noise. Images are divided in user-defined small boxes to
allow for sky background changes along the image. Inside these
subsets the mode of the pixel values is assigned to the local
background, while noise is assigned to standard deviation.
Objects are found in the image following these steps: 1) the
algorithm begins when it finds a pixel N sigma above the local
background (N~1 usually). 2) neighboring pixels are added
if the global S/N ratio is improved after adding them. This test is
made for every surrounding pixel. 3) (optional) neighboring pixels
cannot differ significatively from a gaussian seeing profile among
each other. 4) (opt.) Minimum pixels for an object can be
Star-galaxy separation is performed using the algorithm described in
Sebok (1979; AJ 84, 10). However line search is performed on the
whole catalog to allow for star-like objects like QSO's or
ultracompact galaxies.
Once the objects on the image have been identified and astrometric
solutions have been computed, a cross-correlation may optionally be
performed between direct and OP catalogs to eliminate scratches and
other stuff in the images. Moreover, the final end of this
correlation is to allow us to use an optimal spectra extraction. We
also attempt to search emission lines in unpaired objects from the
OP image for the case of very strong lines and weak continuum
objects in the detection limit.
Click here if you want to see one of the candidates selected and the amount of information that OPERA is able to provide. ReferencesAlonso O., Zamorano J., Rego M., Gallego J. 1995, A&AS, 113, 399García-Dabó, C.E., Gallego, J., 1998 A&SS in press |