Cross-matching the Hubble Legacy Archive

Abstract. The Hubble Legacy Archive (HLA) is composed of hundreds of thousands of images taken by a number of different instruments all across the sky. A large number of these images are of overlapping fields and therefore contain many of the same objects. Through the use of a probabilistic cross-matching technique, it is possible to use these overlapping fields to determine a linkage between similar objects. This linkage can then be used to improve astrometry, as well as provide a means for connecting images taken via different instruments. We present a complete description of this algorithm including the cross-matching technique and solution for correcting the astrometry based upon a single 3D rotation. We then test this algorithm upon two different sets of overlapping Advanced Camera for Surveys (ACS) data, one with a common filter and one with differing filters, to great success. In both the single filter and different filter cases, a cross-match was performed and an astrometric correction possible in which the relative astrometry was improved from the 0.12” inherent to the ACS images in HLA to 0.0034” and 0.0020”, respectively. These values are well into the subpixel level.