Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.stsci.edu/ftp/science/goods.save/Internal/HST/Products/Color-Images/
Дата изменения: Fri Jul 25 17:54:02 2003
Дата индексирования: Mon Dec 24 00:57:20 2007
Кодировка:

Поисковые слова: arp 220
Full-Color Mosaic Pictures for CDF-S and HDF-N GOODS: The Great Observatories Origins Deep Survey

Full-Color Mosaic Pictures for CDF-S and HDF-N

Anton Koekemoer & Zolt Levay

Introduction

Here we present full-color images for the first 3 epochs of both the CDF-S and HDF-N, based on the combined mosaic fits files for the b,v,i and z bands. The fits file for each band was initially binned down by a factor of 2 to create a 9000x12000-pixel image (with 0.1 arcsecond pixels), and converted to 8 bits using an optimal sampling of the dynamic range. The 4 bands for each of the fields were then loaded into Adobe Photoshop and combined as four separate colour layers, before being written out as an RGB TIF file.

These are the images that were displayed at the June 2003 AAS meeting. The downsampling by a factor of 2 is primarily to facilitate working with the RGB file, since the full mosaic is impractical to manipulate in most of the standard image display sotware (eg photoshop, xv). Smaller pieces of the mosaic at the full 1-to-1 pixel scale can readily be generated upon request.

Full-scale tiff files:

CDFS: cdfs123-bviz.tif (143 Mb)

HDFN: hdfn123-bviz.tif (145 Mb)

Further Details

The process of producing combined mosaics for each epoch begins with the fully reduced single-epoch tiles, as produced by Anton's MultiDrizzle script running in Harry's automated GOODS pipeline. These are the tiles that are currently available as the GOODS v0.5 science products. Catalogs are created for each of these single-epoch tiles, using SExtractor with parameters set to produce optimal astrometry for point sources. The catalogs for all the tiles, from all the epochs to be combined, for a given ACS filter band, are then matched to the catalog from the ground-based WFI-R image in a single step, in a technique developed by Stefano Casertano, to derive a fully self-consistent astrometric solution for all the tiles simultaneously. These solutions are described in terms of the 8-parameter World Coordinate System (WCS) keywords for each tile, which include the x and y location of the reference pixel on the chip (CRPIX1,CRPIX2), the Right Ascension and Declination of the reference pixel (CRVAL1,CRVAL2), and the 4-valued CD matrix describing the transformation from detector to sky coordinates (CD1_1, CD1_2, CD2_1, CD2_2).

The header WCS of each tile is then updated with the newly derived values from the solution, and is subsequently propagated back to the original input ACS "flt.fits" files that were used to create each tile, using a script called "backwcs" written by Richard Hook. These flt.fits files are then directly drizzled, using Richard's WCS-based "wdrizzle" program, onto a series of relatively shallow output mosaics where each mosaic has a depth corresponding to one exposure (typically about half an orbit). Thus, for example, a 3-epoch z-band dataset will at this point be drizzled onto 12 separate mosaics, 4 per epoch (corresponding to the 4 flt.fits files for each tile in that epoch), with each mosaic containing only a single flt.fits file per tile location. On the other hand, a 3-epoch v-band (or i-band) dataset will be drizzled onto only 6 separate mosaics because in these bands each tile has only 2 exposures per epoch.

Once the set of shallow single-exposure mosaics has been created, these are then combined to create a clean "median" mosaic, using signal-to-noise thresholds to reject cosmic rays and any remaining bad pixels. This technique is extremely robust at producing a clean median mosaic, with the primary requirement being accurate astrometric alignment between epochs. The relevant pieces of this median mosaic are then transformed back to the frame of each of the original input flt.fits files, using "wblot" which is essentially a WCS-based version of "blot". Next, the standard dither package tasks of "deriv" and "driz_cr" are used to compare this blotted image and its derivative image with the original input flt.fits file, and generate a cosmic ray mask based on the comparison. Finally, all the flt.fits files, together with their newly created cosmic ray masks, are drizzled using "wdrizzle" onto a single output mosaic, which has units of countrate in electrons/second in each pixel.

A weight file is also created for each mosaic, which corresponds to the inverse variance of each pixel. Initially an inverse variance image is created for each input flt.fits file, taking into account the measured sky background values, along with the different detector gains and readnoise values for each amplifier on the ACS/WFC chip, and the pixels that have been flagged as being bad. These inverse variance images are then additively combined in the final step when the mosaic is created.