Recovering 3D structural properties of galaxies from SDSS-like photometry

Because of the 3D nature of galaxies, an algorithm for constructing spatial density distribution models of galaxies on the basis of galaxy images has many advantages over surface density distribution approximations. We present a method for deriving spatial structure and overall parameters of galaxies from images and estimate its accuracy and derived parameter degeneracies on a sample of idealised model galaxies. The test galaxies consist of a disc-like component and a spheroidal component with varying proportions and properties. Both components are assumed to be axially symmetric and coplanar. We simulate these test galaxies as if observed in the SDSS project through ugriz filters, thus gaining a set of realistically imperfect images of galaxies with known intrinsic properties. These artificial SDSS galaxies were thereafter remodelled by approximating the surface brightness distribution with a 2D projection of a bulge+disc spatial distribution model and the restored parameters were compared to the initial ones. Down to the r-band limiting magnitude 18, errors of the restored integral luminosities and colour indices remain within 0.05 mag and errors of the luminosities of individual components within 0.2 mag. Accuracy of the restored bulge-to-disc ratios (B/D) is within 40% in most cases, and becomes worse for galaxies with low B/D, but the general balance between bulges and discs is not shifted systematically. Assuming that the intrinsic disc axial ratio is < 0.3, the inclination angles can be estimated with errors < 5deg for most of the galaxies with B/D < 2 and with errors < 15deg up to B/D = 6. Errors of the recovered sizes of the galactic components are below 10% in most cases. In general, models of disc components are more accurate than models of spheroidal components for geometrical reasons.