HDF source detection and photometry

The catalogs released on January 15 had an error in the x,y and RA, DEC coordinates. They have been corrected and re-released January 16. The new versions will be stored as FITS tables (*.fits) rather than stsdas tables (*.tab), but may not propagate to the ftp site until mid-day on January 16.

In any event, the catalogs are released with a huge caveat emptor. We expect to replace these preliminary catalogs with more trustworthy ones within a few weeks. We would also be happy to archive any catalogs that others produce with different software or source detection algorithms.

The catalogs are available as FITS files of stsdas tables by anonymous ftp from stdatu.stsci.edu.

A brief description of the catalogs can be found in the README file.

The source detection and photometry were carried out with a revised version of FOCAS (Tyson and Jarvis 1979), created by Kurt Adelberger and Chuck Steidel at Caltech. This program is still in the testing phase, but has many improvements over previous versions. The program works by smoothing the data with a fixed kernel, typically the width of the PSF, and then searching for pixels more than a certain number of sigma above the local sky background.

Our source detection and photometry were carried out on the drizzled images. Source detection was done using the sum of the F606W and F814W images to provide the maximum limiting depth. The kernel was an HST psf derived from a star in the field. The minimum area required for detection was 16 continguous drizzled above (0.04 arcsec) above the detection threshold. For the catalog presented here, we have set the threshold fairly conservatively at 5-sigma.


These catalogs were made at the very last minute before the AAS meeting. We recommend that you use them only as a comparison to your experiments with object detection. A visual inspection suggests that the catalog is not too bad -- the main problem seems to be oversplitting in general, and big clouds of spurious objects surrounding bright galaxies. It is possible that some fine tuning of the FOCAS "significance" parameter can help in avoiding these things. In isolated areas FOCAS appears to be finding most of the sources and not introducing too many spurious ones.

FOCAS detection and photometry rely heavily on detailed knowledge of the noise properties of the image. The drizzling process introduces correlated noise between pixels that is not yet fully understood. Therefore the formal significance of the detections and the uncertainties on the magnitudes must be taken with a grain of salt.

Copyright © 1997 The Association of Universities for Research in Astronomy, Inc. All Rights Reserved.

Harry Ferguson ferguson@stsci.edu 1/14/96