Automatic analysis of crowded fields

Abstract

There are many important two-dimensional data-reduction problems in astronomy that are not amenable to conventional automatic analysis. Typically these problem areas arise in fields where the number density of images is high and where the local sky background may vary rapidly. Within such regions the image number density becomes so high that the majority of images overlap, even at relatively high isophotes, and simple image-parameter estimation algorithms become confused. In this paper the potential for a fully automatic method that is both robust and efficient in terms of computer requirements, capable of dealing with complex multiple overlaps, and able to generate the optimum estimates of image parameters is examined. By applying the theory of maximum-likelihood parameter estimation, a coherent strategy is devised which leads to the development of a fully automatic system capable of producing reliable results toward the center of globular clusters and within the dense regions of nearby resolved galaxies.