[Previous] | [Session 70] | [Next]
R.A.M. Walterbos, D.A. Thilker (NMSU), R. Braun (NFRA), V. Fierro (NMSU)
Automated determination of the positions, fluxes, and sizes of HII regions in galaxies is crucial for efficient and reproducible characterization of their star formation properties. Advantages of such an approach over conventional measurement by visual inspection, other than the obvious increase in speed, are that the effects of reduced spatial resolution for more distant galaxies and of differences in limiting sensitivity can be quantitatively investigated. In crowded environments it is practically impossible to cleanly separate the contribution of neighboring extended objects to the observed surface brightness of a given pixel. Correcting for source overlap the way DAOPHOT does for stellar crowded field photometry would require models describing the structure of all types of extended HII regions in nearby galaxies, an infeasible approach at present. Consequently, any automated photometric procedure optimized for working on HII regions must allow for tolerance in observed source morphology. This can be accomplished using an iterative approach to "grow" sources from an initial guess at the shape. We have developed such a method, called HIIphot. The program incorporates object recognition techniques in order to properly classify center-brightened and ring-like objects. Our iterative growing procedure then permits departures from idealized seed models used during the initial search step. HIIphot carefully controls growth in the saddle points between neighboring regions and can therefore be effective at recovering all detected flux (within the unavoidable constraints of crowding) rather than only the contribution from each region's brightest pixels. Corrections for the diffuse background emission are computed carefully by incorporating a low-order, surface fit constrained by the intensities in local background annuli which exclude neighboring objects. We present results for 10 nearby galaxies, including HII region luminosity functions going an order of magnitude deeper than earlier work.