When you visualize a fits file with personalized min/max value, FITS keywords DATAMIN and DATAMAX are changed in file concerned. But DATAMIN and DATA MAX are fits standards keywords https://fits.gsfc.nasa.gov/fits_standard.html KEYWORD: DATAMAX REFERENCE: FITS Standard STATUS: reserved HDU: image VALUE: real COMMENT: maximum data value DEFINITION: The value field shall always contain a floating point number, regardless of the value of BITPIX. This number shall give the maximum valid physical value represented by the array, exclusive of any special values. So DATAMIN and DATAMAX can't be used to store theses values as it correspond to camera numeric depth. (for example, for a 16 nits camera, value are DATAMIN = 0 and DATAMAX = 65535 ) For example, Audela store it in MIPS-MAX mans MIPS-MIN keywords. (not in standards...) May be Kstars can do the same... or store it in other non-standards keywords. Comments to read from this very bad placed bug from TallFurryMan https://github.com/indilib/indi/issues/693 I agree, DATAMIN/DATAMAX do not relate to the actual minimum/maximum values of FITS data, but indicate what the dynamic range of the physical equivalent of the FITS data is. So they can't be changed at all. Not that I trust particularly that plug-in, but here's a side reference on the GIMP plug-in reading FITS files: https://gitlab.gnome.org/GNOME/gimp/blob/master/plug-ins/file-fits/. That plug-in didn't change much since 1999... FITS header is read in https://gitlab.gnome.org/GNOME/gimp/blob/master/plug-ins/file-fits/fits-io.c. There are two modes, stated "automatic" and "by DATAMIN/DATAMAX": With "automatic", FITS data is applied a reference change from the minimum/maximum value of FITS data to the [0,DATAMAX] interval. This is just weird as an automatic mode. With "by DATAMIN/DATAMAX", FITS data is applied a reference change from origin BZERO and scale BSCALE. The minimum/maximum value of FITS data is considered to be the inverse reference change of DATAMIN/DATAMAX. This is proper conversion to physical data values. ------------