extracting wavelength calibration from .FITS

extracting wavelength calibration from .FITS

Craig (guest) 18 May 2010 06:02

I've been reducing data with VIPGI, which I have now exported to a FITS datacube. When I open the cube with a program like QFitsView, the spectrum for each spaxel is displayed along the bottom of the window when I mouse over the corresponding region of the image, along with the axis scale calibrated to wavelength. Obviously then, the wavelength calibration must be stored somewhere in the FITS header or something.

However, I am performing most of my data analysis and manipulation in IDL and Matlab. Importing the datacubes with the standard fitsread functions does not automatically recognise this calibration, and when I plot spectra, they are calibrated to pixel number. I pretty much expected that this would be the case, but my question is, how does one extract the wavelength calibration from the FITS file? Are there any tools to do this in IDL for datacubes? Failing that, where abouts in the FITS header is this information stored?

Any help or info would be most appreciated…

Re: extracting wavelength calibration from .FITS

Katrina (guest) 18 May 2010 08:41

I don't have vipgi cubes in front of me, but the wavelength information, along with the spatial information, is usually stored in the fits header entries CDELT1/2/3, CRVAL1/2/3. The first is the dispersion, the second the starting wavelength. There may also be something including "PIX1/2/3", which indicates to which pixel the CRVAL refers (usually 0/1 but sometimes not).
I'm curious - our vimos cubes don't display the wavelength scale in QFITSVIEW, can you tell me what CDELT1/2/3 and CRVAL1/2/3 values you have in your fits header? I am wondering if our cubes are just the "wrong way around", so-to-speak.
I have always just created the wavelength array in IDL from these information, I am not sure if any of the fits readers of IDL do this. It would make sense if there were one!

Re: extracting wavelength calibration from .FITS

Becky (guest) 30 Jul 2010 16:10

Hi Katrina,

I was wondering if you managed to fix the wavelength scale problem in Qfitsview with VIMOS data? I'm having the same problem, I usually just use idl to look at the cubes but Qfitsview seems like a handy tool for a quick look at the data.

Cheers,

Becky

Re: extracting wavelength calibration from .FITS

katrina (guest) 03 Aug 2010 16:06

I didn't try very hard to get it to work. I am wondering whether it is because qfitsview expect the wavelength array to be a particular one of axis 1 2 or 3, but I have not looked into the code to see. I'll have another poke about later this week

Re: extracting wavelength calibration from .FITS

katrina (guest) 03 Aug 2010 16:52

Ha,

so, in iraf I simply added in/made sure were there (with "ehead") the following fits-headers
CDELT1/2/3 (dispersion)
CRVAL1/2/3 (1st value)
CD1_1/2/3 (dispersion also)
CTYPE1/2/3 (being 'RA', 'DEC', 'Lambda')
and qfitsview is happy with that
I don't know which one(s) did the job, but all together certainly did
my data have axes of RA, Dec, wavelength, in that order. so if doing this on other cubes does not work, then I guess maybe the cubes have to be organised in that order

Re: extracting wavelength calibration from .FITS

Becky (guest) 19 Aug 2010 09:48

Heya,

Thanks very much. Worked a treat!