[fitswcs] WCS for co-moving maps

[Steve Allen]

On Thu 2015-09-17T13:47:41 +0200, Ian Avruch hath writ:
> The situation is that we are making maps of the sky around
> moving solar system objects (SSOs), such as comets.  The telescope
> slews in a pattern while tracking the apparent position of the SSO,
> so that while on the sky the sample points are sparse, when reduced
> to offsets relative to the SSO the sample points fill an area.
>
> Neighboring points in the offset map differ by fixed increments
> of angular distance parallel to RA or Dec (ICRS) from the
> _instantaneous_ position of the SSO at the different sample times.
> So they are not increments along RA or Dec; this should be clear.

This shares some characteristics with one of our instruments.
In our case we also determined that the FITS WCS formalisms
cannot describe the dataset.  Instead we took a shortcut.
But if I understand this SSO tracking then there may be a
way of accurately describing these images while still displaying
them in a straightforward fashion.

Our application is on the 2.4 m Automated Planet Finder az-el optical
telescope which uses a CCD guide camera to keep the target on the slit
of a spectrometer.  To assist with analysis of quality issues in the
spectral data we can save the entire sequence of guide camera images
during an exposure as a movie in a 3-d FITS file where the third
axis is time.

There is no image rotator, so the cardinal directions of the sky
rotate during the movie.  The WCS formalism provides no way to
alter the axes in a functional fashion.  Therefore no single
WCS can describe the celestial WCS for the entire movie.

Our shortcut is to write two alternate WCS into the header.
One WCS is spatially correct for the initial image in the stack.
One WCS is spatially correct for the final image in the stack.
Both are temporally correct for the entire stack.

If I understand this SSO application there are many small 2-d images
which are taken at different times in different directions.  If these
were stored as separate IMAGE extensions then each one could have
multiple alternate WCS in them.  One of those WCS could be the offset
coordinates for the particular frame.

If, say, the WCS for the co-moving frame were alternate version M, and
if such a multi-extension FITS file was to be displayed using the
SAOimage ds9 application, and if ds9 were instructed to display the
IMAGE extensions using alternate WCS M then ds9 would present a view
of the various pieces of the image all tiled in a manner around the
moving SSO.

Other alternate WCS versions in each IMAGE extension could give
the instantaneous ICRS coordinates, detector coordinates, etc.

--
Steve Allen                 <sla at ucolick.org>               WGS-84 (GPS)
UCO/Lick Observatory--ISB   Natural Sciences II, Room 165   Lat  +36.99855
1156 High Street            Voice: +1 831 459 3046          Lng -122.06015
Santa Cruz, CA 95064        http://www.ucolick.org/~sla/    Hgt +250 m