Extremely large pixel values
fergusjm wrote on Jul 09, 2015
Hello, I'm having a problem after dividing my trimmed frames by the response to the combined flat field.
# IMAGE NPIX MEAN STDDEV MIN MAX
trim_26_1354 1890304 848.7 1204. -2591. 49869.
# IMAGE NPIX MEAN STDDEV MIN MAX
flat_hk_nov28_2013.fits 1890304 0.9995 0.08823 0. 2.675
imarith trim_25_1354 / flat_hk_nov28_2013.fits sdss_25
# IMAGE NPIX MEAN STDDEV MIN MAX
sdss_25 1890304 2.703E14 7.193E16 -3998. 7.428E19
How does the mean jump to 2.703E14?
Mike Fitzpatrick wrote on Jul 09, 2015
Could you upload the two test images to the anonftp at ftp://iraf.noao.edu/pub so we can try to reproduce the problem? Also, what version of IRAF is this and on which platform (Mac/Linux, 32- or 64-bit)??
fergusjm wrote on Jul 09, 2015
When I try to access the pub, it won't let me drag and drop anything into the folder. I'm not exactly sure where the anonftp is, once I log onto the server I get this http://i.imgur.com/R0RuzHY.png window.
Mike Fitzpatrick wrote on Jul 09, 2015
From a Terminal window, do:
% ftp -A iraf.noao.edu
..... login as 'anonymous', give your email as the password
ftp> cd pub
ftp> bin # need to set binary mode
ftp> put foo.fits # upload file, repeat as needed
ftp> quit
fergusjm wrote on Jul 09, 2015
I believe they were uploaded succesfully. The computer is a mac and I'm running IRAF with xgterm. I am using IRAF V2.16 updated on 2015-3-24.
Mike Fitzpatrick wrote on Jul 09, 2015
Your Flat file has several pixels with values such as 1E-16 which is skewing the output. You can use the IMREPLACE command to fix these pixels to some known value you like better, e.g.
cl> imreplace flat_hk_nov28_2013.fits 0.0 upper=0.01
would replace all values less than 0.01 with 0.0. The resulting IMSTAT looks better but depending on your particular application you may want to replace values in the output as well.
Last post on Jul 09, 2015