To perform smooth pipeline operations from data of various instruments and telescopes, we need to specify a set of header keywords for FITS images that enter THELI processing. In particular, the header needs to contain all necessary information for astrometric and photometric calibration of data. To minimise possible confusions we substitute the original image header with one that contains ONLY the minimum set of required information. A header from a raw image that you want to process with THELI typically looks like:
SIMPLE = T
BITPIX = 16
NAXIS = 2
NAXIS1 = 2142
NAXIS2 = 4128
BSCALE = 1.000000
BZERO = 32768.000000
CTYPE1 = 'RA---TAN '
CTYPE2 = 'DEC--TAN '
CRPIX1 = -412.00
CRPIX2 = 3888.00
CD1_1 = -6.610000E-05
CD2_2 = 6.610000E-05
CD2_1 = 0.000000E+00
CD1_2 = 0.000000E+00
CRVAL1 = 170.622467
CRVAL2 = -21.721230
RADECSYS= 'FK5 '
EQUINOX = 2000.00
FILTER = 'BB#Rc/162_ESO844 '
OBJECT = 'DEEP3b '
AIRMASS = 1.107282
EXPTIME = 299.92
IMAGEID = 3
GABODSID= 726
ZP = -1.000000
COEFF = 1.000000
DUMMY1 = 0
DUMMY2 = 0
DUMMY3 = 0
DUMMY4 = 0
.
.
We note that all THELI processing happens on single-chip images - other FITS formats such as Multi-Extension-FITS files are never used. If your initial data come in some FITS-cube format, you need to split them first. This step typically also takes care of setting up the required image headers. The THELI distribution contains a C-program mefsplit which performs these tasks. For specfic instruments it is typically wrapped by a shell script. See process_split_WFI_eclipse.sh (for WFI@MPG/ESO2.2m) and process_split_MEGAPRIME_eclipse.sh (for MEGAPRIME@CFHT) for examples.
Note