DrizzlePac

DrizzlePac
Name	DrizzlePac
Developer	Space Telescope Science Institute
Released	2000s
Programming language	Python, C
Operating system	macOS, Linux, Windows
Genre	Astronomical image processing
License	Open source / STScI software

DrizzlePac is a software package for combining and optimizing astronomical imaging data, designed primarily for data from the Hubble Space Telescope and similar observatories. It provides tools to align, drizzle, and clean exposures, integrating algorithms for image registration, distortion correction, cosmic ray rejection, and PSF handling. DrizzlePac is widely used within projects at institutions such as the Space Telescope Science Institute, the European Space Agency, and research groups at universities and observatories worldwide.

Overview

DrizzlePac implements image combination techniques built on the Drizzle algorithm and associated utilities, enabling astronomers to produce high-fidelity mosaics, deep stacks, and resampled images from instruments like Wide Field Camera 3, Advanced Camera for Surveys, and Wide Field Planetary Camera 2. Users rely on DrizzlePac alongside software such as IRAF, AstroDrizzle, Astropy, SExtractor, and DAOPHOT to perform tasks from astrometric alignment to cosmic-ray cleaning. The package interoperates with catalog services and reference frames including Gaia, 2MASS, Sloan Digital Sky Survey, Pan-STARRS, and UCAC4 to improve absolute and relative astrometry.

History and Development

Development traces to algorithmic work on the original Drizzle method, first employed for projects like the Hubble Deep Field and later adapted for instruments deployed during servicing missions to the Hubble Space Telescope. Core development and maintenance have been led by teams at the Space Telescope Science Institute with contributions from staff at the European Southern Observatory, STScI affiliates, and collaborators at research centers such as Harvard–Smithsonian Center for Astrophysics, California Institute of Technology, Massachusetts Institute of Technology, and Johns Hopkins University. The project evolved through iterations responding to the needs of survey programs like CANDELS, COSMOS, PHAT, and legacy programs tied to the Hubble Legacy Archive and mission calibration efforts. Integration with community tools like GitHub and adoption by consortia including HST Treasury Programs have shaped release cycles and user support.

Features and Components

DrizzlePac comprises multiple modules and utilities: image registration tools, the primary drizzle combination engine, cosmic-ray rejection routines, and image inspection and manipulation utilities. Its registration utilities integrate with catalogs from Gaia DR2, Hipparcos, USNO-B1.0, and survey catalogs from CFHT, Subaru Telescope, and VISTA for reference matching. The drizzle engine supports parameters for pixel fraction, drop size, output pixel scale, and weight map handling, aligning with instruments such as ACS, WFC3/UVIS, and WFC3/IR. Ancillary components include tasks for distortion models based on instrument calibration files maintained by teams at STScI Calibration Group and interaction scripts compatible with environments like Python and PyRAF. Visualization and quality-assurance tools interoperate with viewers like DS9, reduction packages such as IRAF/STSDAS, and analysis suites including TOPCAT and APLpy.

Workflow and Usage

A typical DrizzlePac workflow begins with retrieval of raw and calibrated exposures from archives such as the Mikulski Archive for Space Telescopes or project-specific repositories, followed by preprocessing with calibration pipelines maintained by STScI or observatory pipelines for instruments like Keck or Gemini Observatory. Users perform source detection using SExtractor or DAOStarFinder to generate input catalogs, match against external astrometric references like Gaia or Pan-STARRS1, and refine offsets with DrizzlePac registration tasks. After alignment, the drizzle combination step creates final mosaics, often followed by PSF homogenization using models derived from TinyTim, empirical stacks, or tools developed at Space Telescope Science Institute. Workflows are adapted into scripts and pipelines for survey-scale processing employed by teams at institutions such as University of California, Berkeley, Princeton University, University of Cambridge, and Max Planck Institute for Astronomy.

Performance and Limitations

DrizzlePac scales effectively for modest ensembles of HST exposures but encounters computational and I/O limits when applied to extremely large mosaics or wide-field mosaics from facilities like VISTA or wide-area surveys such as LSST. Performance depends on factors including memory bandwidth, CPU parallelism, and available storage; users often deploy processing on clusters operated by centers like NASA Ames Research Center, National Center for Supercomputing Applications, or university HPC clusters. Limitations include sensitivity to input astrometric errors from catalogs such as older releases of USNO or PPMXL, challenges with undersampled PSFs for instruments operating near the Nyquist limit, and residual systematics when combining data with heterogeneous background levels from programs like CANDELS and GOODS. Ongoing development addresses GPU acceleration, tighter integration with Astropy data models, and improved handling of large mosaics.

Applications and Notable Results

DrizzlePac has been instrumental in producing high-resolution, deep-field images used in landmark results from programs such as the Hubble Ultra Deep Field, CANDELS, Frontier Fields, and targeted studies of gravitational lenses in surveys including work by teams at University of Oxford, Yale University, and University of Tokyo. Outputs processed with DrizzlePac feed analyses for galaxy morphology studies, weak lensing measurements in programs tied to CFHTLenS, stellar population work in nearby galaxies like those studied by the ANGST team, and exoplanet transit imaging from coordinated campaigns with facilities such as Spitzer Space Telescope and Kepler. Community adoption spans research groups at Stanford University, University of Chicago, Princeton, Carnegie Observatories, and observatories including NOAO and European Southern Observatory, underpinning hundreds of peer-reviewed publications and archival products in major data releases.

Category:Astronomical image processing software