The High Dispersion Background Algorithm in NEWSIPS

Previous abstract Next abstract

Session 45 -- IUE Final Archive
Display presentation, Thursday, January 13, 9:30-6:45, Salons I/II Room (Crystal Gateway)

[45.08] The High Dispersion Background Algorithm in NEWSIPS

M.A.Smith (CSC), C. A. Grady (ARC), P. O'Brien (UCL), M. DelaPena (CSC), J. Nichols (CSC), M. Garhart (CSC), B. Coulter (CSC), A. Michalitsianos (GSFC)

A two-dimensional interpolating scheme, followed by modeling of the point spread function, is outlined for use in the final archiving NEWSIPS program in removing background fluxes of high dispersion IUE images. So far our tests have been limited mainly to SWP camera images.

An integral facet of our background removal algorithm, basisiue , is its execution in a totally automated environment. Toward this end several conditioning steps are required before the background fluxes can be sampled. These include the removal of ``wiggles" of echelle orders as well as rotation of the camera format and removal of order ``splaying" and avoiding pixels with high fluxes due to permanent image blemishes and cosmic ray hits. Image-specific pixels with such pathologies are eliminated, along with on-order pixels, for a sample of pixels along 26 "swaths" (SWP camera) in the cross-dispersion direction. Smoothed, one-dimensional 7-th degree Chebyshev fits are then computed from the interpolated fluxes modified by a global point spread function determined from the interorder overlap pattern in an ensemble of science images. A second set of continuous Chebyshev functions, perpendicular to the first, is computed next along the positions of the IUE orders by interpolating across fluxes determined from the first set. Thus, this algorithm determines both the of background fluxes at arbitrary locations on the image and also determines the amount of interorder flux-overlap among short-wavelength orders, which is necessary to the final extraction of spectral fluxes.

This work has been supported under NASA Contact NAS5-31230 to the Computer Sciences Corporation.

Thursday program listing