selecting images by quality?
astromaus wrote on Aug 05, 2010
Hello,
I'm new here and new to IRAF.
I would like to do some photometry measurements (semi-professional). since I'm a physisist but not an astronomer, I need/want to work myself through IRAF (I used it for a course in university a few years ago and I liked it).
anyway...
Due to very bad seeing and some wind disturbing our little telescope,
I have a bunch of bad pictures among the good ones.
is there a package in IRAF that can sort out bad pictures (oval stars, not sharp and so on)?
thanks
astromaus
I'm new here and new to IRAF.
I would like to do some photometry measurements (semi-professional). since I'm a physisist but not an astronomer, I need/want to work myself through IRAF (I used it for a course in university a few years ago and I liked it).
anyway...
Due to very bad seeing and some wind disturbing our little telescope,
I have a bunch of bad pictures among the good ones.
is there a package in IRAF that can sort out bad pictures (oval stars, not sharp and so on)?
thanks
astromaus
Mike Fitzpatrick wrote on Aug 05, 2010
'Image quality' is very subjective, there's nothing to automatically do this by whatever definition you give that term.
That said, a task like STARFIND/DAOFIND will detect objects and produce shape parameters you can use as a criteria for a specific file. Tasks like FIELDS and AVERAGE can be used in a script to process the output of the STARFIND table to see if some average roundness meets your threshold of 'good' vs 'bad', and HSELECT can be used to select by image if there is a header parameter with the info you need. I think you may need to write your own script to do the selection, see the Script Guide under the 'Docs' link on the banner bar to get started and post back if you still have questions.
-Mike
That said, a task like STARFIND/DAOFIND will detect objects and produce shape parameters you can use as a criteria for a specific file. Tasks like FIELDS and AVERAGE can be used in a script to process the output of the STARFIND table to see if some average roundness meets your threshold of 'good' vs 'bad', and HSELECT can be used to select by image if there is a header parameter with the info you need. I think you may need to write your own script to do the selection, see the Script Guide under the 'Docs' link on the banner bar to get started and post back if you still have questions.
-Mike
astromaus wrote on Aug 05, 2010
thank you for your answer :)
then I have to think of something.
In another software tool (ImageJ+astrotools), I calibrated a chi^2-value of the seeing profile/radial plot of the stars in my pictures.
then I selected by this value. the "good" pictures had a chi^2 around 1 and the bad ones above this value.
is there a way to get the chi^2 value of the radial profile in iraf?
and can I somehow influence the way iraf makes the fit of the profile?
because the fitted profiles I've seen from may data so far, didn't look to
good.
then I have to think of something.
In another software tool (ImageJ+astrotools), I calibrated a chi^2-value of the seeing profile/radial plot of the stars in my pictures.
then I selected by this value. the "good" pictures had a chi^2 around 1 and the bad ones above this value.
is there a way to get the chi^2 value of the radial profile in iraf?
and can I somehow influence the way iraf makes the fit of the profile?
because the fitted profiles I've seen from may data so far, didn't look to
good.
Mike Fitzpatrick wrote on Aug 05, 2010
See the RADPROF task in the APPHOT package, there are also radial profile tools available in the IMEXAMINE task. RADPROF might be trickier to set the parameters but has a batch mode so you can process lots of images. I don't think either produce exactly a chi^2 value but do have some sort of fitting error term.
astromaus wrote on Aug 05, 2010
I "flew" over the functions of radprof.
when I tried it with one picture in "interactive mode" I get:
I guess that's the same problem as here?:
http://iraf.net/faqman/index.php?op=view&t=140
is there any way to actually plot the radial profile with radprof or do
I have to plot it with different programs like gnuplot?
But I guess, all this is going to be too complicated, so I will stick
with ImageJ+astrotools in this case. This generates a list of pictures which I can use in IRAF to analyse. :-)
for future measurements I will see to it, that I'll get better pictures :D
thank you, for your help
when I tried it with one picture in "interactive mode" I get:
Warning: Graphics overlay not available for display device.
I guess that's the same problem as here?:
http://iraf.net/faqman/index.php?op=view&t=140
is there any way to actually plot the radial profile with radprof or do
I have to plot it with different programs like gnuplot?
But I guess, all this is going to be too complicated, so I will stick
with ImageJ+astrotools in this case. This generates a list of pictures which I can use in IRAF to analyse. :-)
for future measurements I will see to it, that I'll get better pictures :D
thank you, for your help
Last post on Aug 05, 2010