WCSFixer - A Web-enabled

1
THE US NATIONAL VIRTUAL OBSERVATORY
WCSFixer - A Web-enabled IRAF Plate Solver
Michael Fitzpatrick NOAO
2
Introduction

• Final-year planning for the NVO by its
  Science Steering Committee identified a number of
  capabilities desirable as end-user applications.
  The ability to automatically correct (or
  determine) the astrometric solution of image data
  ranked as a high-priority task. Much of the
  needed functionality already existed in IRAF, but
  some level of interactive use is needed since no
  automated solution can be completely robust for
  all available VO data.
• The challenge then was to implement the desktop
  reduction experience in the web browser while
  still providing a batch mode for processing large
  numbers of images as web services or from client
  code.

3
Implementation

• - CGI/POST web interface
• Does form processing, some interface elements
• - IRAF CL scripts to implement functionality
• !cl script to define IRAF environment prior to
  calling standard task
• Same scripts run unchanged in IRAF desktop env
• Host tools (e.g. STILTS, Sextractor available,
  image conversion tools)
• - Embedded JavaScript output to drive interface
• - CSS to aid layout/appearance

4
Real-World Data

• Example Problems encountered (and solved)
• DSS images have no WCS keywords
• CFHT data contain a 64-bit integer id as a
  keyword (non-standard)
• Axis mappings (RA/Long not always first axis)
• WCS Projections not handled by IRAF
• X-ray data not suited to Sextractor-like object
  detection
• Star trails and focus problems
• Junk Strip in WFPC data
• False solutions -- rms of fit not a good
  indicator
• Large parameter space
• Legacy keywords
• Use of RA/DEC not standard
• specific IRAF keywords (e.g. DATASEC) no longer
  valid

5
Spin-off Applications/Services

• Image catalog generation/crossmatch
• WCS Summary
• Preview image
• WCS manipulation
• Tweak existing WCS (from test app)
• Swap axes
• Change projection/epoch
• Multiple deployments
• Web application
• SOAP/WSDL web service
• Desktop app in IRAF environment

6
Lessons Learned

• Problem itself is a rich area of research
• Labor-intensive way to develop
• In part because of this particular problem
• Re-use in other deployments a bit if an offset
• New appreciation for data quality concerns in VO
• Legacy techniques and tools can still produce
  useful VO applications