[AstroPy] A Mystery: DAOStarFinder Not Finding an Obvious Star in Just One Image

[Larry Bradley]

Bob,

As Eric suggested, it could be rejected because of the sharpness,
roundness, or peakmax criteria.  You could test that by relaxing those
bounds.

The other reason it could be rejected is because the threshold is too
high.  Actually, it's the combination of threshold and fwhm that matter
here because the thresholding is performed on the convolved image (e.g., as
you increase the fwhm, you generally need to lower the threshold).  The
first thing I would try here is lowering the threshold.

Larry

On Sat, Feb 26, 2022 at 3:30 PM Eric Jensen <ejensen1 at swarthmore.edu> wrote:

> Hi Bob,
>
> You might take a look at the individual pixel values for that star in that
> particular image - I wonder if it might have a cosmic ray hit or some other
> anomaly in the pixel values that might cause it to be rejected, for example
> by exceeding the “peakmax” setting, or by throwing off the roundness or
> sharpness values.
>
> Just a thought -
>
> Eric
>
>
> > On Feb 26, 2022, at 2:56 PM, Robert Carr <rspencercarr at gmail.com> wrote:
> >
> > I am hoping for help with a mystery!
> >
> > I am mystified by a problem with DAOStarFinder.  DAOStarFinder has
> seemed to be working flawlessly.  But a couple days ago I was surprised to
> discover that it refused to find a target in one of my images.  The target
> is bright: 15.3 magnitude in g with a peak flux of about 350.  The star's
> image is quite round.  DAOStarFinder has no problem finding other stars of
> comparable luminosity, as well as a great number of brighter and dimmer
> stars, all the way down to about 38.  DAOStarFinder finds the star in other
> better and worse frames.  But it refuses to find my target in just this one
> image!  Why?!  This is completely maddening.  And worrying: what else might
> it be missing?
> >
> > The relevant code is very simple.  I have set few options for
> DAOStarFinder:
> > >>from photutils import DAOStarFinder
> > >>fwhm = hdulist[0].header['FWHM']*2.5
> > >>source_snr = 1.5
> > >>bkg_sigma = mad_std(data)
> > >>daofind = DAOStarFinder(fwhm=fwhm, threshold=source_snr*bkg_sigma)
> > >>sources = daofind(data)
> >
> > The numbers above work out to be:
> > - fwhm = 4.76 * 2.5 = 11.9
> > - bkg_sigma = 8.9
> > - threshold = 13.3
> >
> > I have tried playing with the above settings with no effect.  I have
> rounded up the usual suspects but the mystery remains.  It is like the star
> has a cloak of invisibility when it comes to DAOStarFinder for just this
> image!
> >
> > Clues and leads on solving my mystery are very much appreciated.  Thank
> you!
> >
> > Bob Carr
>
> _______________________________________________
> AstroPy mailing list
> AstroPy at python.org
> https://mail.python.org/mailman/listinfo/astropy
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mail.python.org/pipermail/astropy/attachments/20220228/0211719d/attachment.html>