[daip] blanked pixels inside fits

[Eric Greisen]

Niruj Mohan Ramanujam writes:

 > I have a question about blanked pixels in FITS files.
 > Documents mention that if BITPIX is -32 or -64 (which is true for most 
 > files these days) and if there are blanked pixels, then the value used 
 > is not the value of the keyword BLANK but the IEEE NAN. Though the value 
 > is usually given as -1.
 > 
 > Can I assume that if there is no BLANK keyword in a FITS file, then 
 > there will be no blanked pixels present ?

Most definitely not.

The standard says:

BLANK keyword: This keyword shall be used only in headers with
positive values of BITPIX (i.e. in arrays with integer data).
...  The value field shall be an integer ...

Later it says that Nan's shall be used in the f.p. data.

So your software must look for Nan's and cannot depend on any header
keyword.  The search for Nan's is not too hard - see $APLGEN
subroutines ZR32RL and ZR64RL.

Eric Greisen