Multi-extension (mosaic) FITS solving

[co...@nau.edu]

I'm hoping I can use Astrometry.net to work on a bunch of multi-extension FITS files (mosaics). 

I saw the --extenesion option, and that multiple extensions are supported for tables, but is there a way to solve the whole frame and output a new fits file?

Any advice would be much appreciated.

[Dustin Lang]

Hmmm, I don't think there is currently an easy way to do this.

The ugly way would be to run solve-field on each extension in turn, outputting to different files, and then use a combination of fitsgetext to extract single extensions, and cat to glue the pieces back into a multi-HDU output.

Yuck.

--dustin

[co...@nau.edu]

Thanks for the idea, I hadn't thought of that path exactly.

For a project I led recently (SAFARI: Searching Asteroids For Activity Revealing Indicators, PASP next issue) I had to split the fields up (one chip to one FITS) anyway because of other tools I was using anyway. Astrometry.net was a life saver for fixing messy WCS headers across ~35K images.

I am doing a much larger (petabyte scale) project now and was hoping to avoid splitting the images because (1) disk usage triples (separate chips + astrometry.net FITS) and (2) distortion analysis, etc. ought to be better across a field instead of a single chip.

I see Astrometry.net docs say it will do multi-extension tables. Do you think I'd be best off feeding Astrometry.net SExtractor output via multi-extension FITS tables and then updating/inserting the headers back into the "original" HDUs?

I'd be happy to share the code if it works, maybe it could become a command-line option.

Thanks again :)

-- Colin