AAS 207th Meeting, 8-12 January 2006
Session 29 Observation Processing, Calibration and Control
Poster, Monday, 9:20am-7:00pm, January 9, 2006, Exhibit Hall

Previous   |   Session 29   |   Next  |   Author Index   |   Block Schedule


[29.03] A Fast Image Cleaning and Coadding Pipeline for the Palomar-QUEST Survey

A. Mahabal, M. Bogosavljevic, C. Donalek, S. G. Djorgovski, R. Williams, M. J. Graham, A. Drake, S. Brunett, J. Lindheim (Caltech), C. Baltay, D. Rabinowitz, A. Bauer, J. Silge, J. Jerke (Yale University), Palomar-QUEST Collaboration

The Palomar-QUEST (PQ) sky survey covers ~ 15,000 sq.deg. on the sky, between Declinations -25 and +25 deg, in multiple passes, in UBIR or riz filters. The survey has been operational for over 2 years, with over 10 TB of data in hand. The synoptic nature of the data as well as stacking images belonging to the same part of the sky lends itself to a variety of science applications including search for transients, high-redshift quasars and other outliers in multi-parameter space.

With these goals in mind, we have been reprocessing the PQ data utilizing a fast c-based pipeline in a VO compatible manner using the grid architecture at Caltech. PQStore is an implementation of VOStore, a VO-standard inventory system to keep track of the different kinds of files by mapping logical names to their physical names across heterogeneous storage devices. It is used to: transfer the incoming ~50 GB data per night to 16 grid nodes, scrub them clean using bias, flat information and extensive clipped-median filtering, insert WCS and transfer the images to the archive. Source extraction is done on the images and the catalogs transferred to a database accessible through a local NVO-standard skynode. Cross-matching allows us to keep track of objects seen multiple times by using virtual ids. This allows for quick access of time history of any object and provides the basis for transient detection with the real time pipeline.

The data are also coadded on to VO-standard HyperAtlas pages. The HyperAtlas paradigm is a new way for federating images which involves reprojecting each image to a common set of pixel planes, then stacking images and detecting sources therein, allowing us to go much fainter.

The fully processed data will be released to the community using NVO-style mechanisms, e.g., a public SkyNode.


The author(s) of this abstract have provided an email address for comments about the abstract: aam@astro.caltech.edu

Previous   |   Session 29   |   Next

Bulletin of the American Astronomical Society, 37 #4
© 2005. The American Astronomical Soceity.