Star catalog for comparison

  • To compare the different image quality measures, I attached the star catalogs (selected_star_cat_11222012.tar.gz and selected_star_cat_11242012.tar.gz) for the exposures we took on 11/22/2012. If you unzip the attached file, you will find many star catalogs, each of which corresponding to one single exposure. Each star catalog file is a multi extension fits file, with each extension corresponds to the extension of the same image.
    • selected_star_cat_11222012.tar.gz contains exposures from 11/22/2012, in which there are exposures with large fwhm due to the tweaking of the hexapod.
    • selected_star_cat_11242012.tar.gz contains exposures from 11/24/2012, in which most of the psfs are small and round.
      By analyzing both of them, we should be able to cover a wide range of PSFs. The corresponding expids of the two star catalogs can be found from the attached files: expid_11222012.txt and expid_11242012.txt
    • if you think there are too many exposures, let's first focus on 155268, 155256, 155309 from 11/24 and 154431, 154442, 154449 from 11/22.
  • To compare the measurement from IH and QR, I would need the fwhm, e1, e2 for each of the stars for each exposure. Here is my suggested format:
    • for each exposure, both IH and QR will produce a file for their measured values, fwhm, e1,e2, flux_radius for each of the star.
    • The output file is a n x m matrix, each row corresponding to one star and the columns are [fwhm, e1, e2, flux_radius]
    • The row should be ordered as follows: 1. in each extension, the order should be the same as the stars in the catalog. 2. the extension runs from 1 to 62 and the appended sequentially.
    • example: the rows correspond to: star1_ext1,star2_ext1..., star1_ext2,star2_ext2, ..., star1_ext62, star2_ext62 ...
  • The column names XWIN_IMAGE and YWIN_IMAGE in the catalog are the location of the stars in the CCD coordinate of that extension. The other columns are as follows, but I do not think you need them for your measurement.
    • 'NUMBER',
    • 'XWIN_IMAGE',
    • 'YWIN_IMAGE',
    • 'X2WIN_IMAGE',
    • 'Y2WIN_IMAGE',
    • 'XYWIN_IMAGE',
    • 'FLUX_RADIUS',
    • 'MAG_AUTO',
    • 'CLASS_STAR',
    • 'FWHM_IMAGE',
    • 'FLAGS'

I have visually checked many of these stars and they are all looks fine. The star selection rules based on these visual examinations are shown by the following function (python):

def firstcutStar(b):
mag = b.MAG_AUTO
flag = b.FLAGS
ok = (mag>=10.5)*(mag<=12)*(flag ==0)*(rad<5.)
radmedian = np.median(rad[ok])
idx = (mag>=10.5)*(mag<=13)*(flag ==0)*(abs(rad-radmedian)<=0.2)
return b[idx]

Here are some examples (the red dots on the plots are the selected stars):

  • R50: the radius that enclose 50% of the total light. As Gary mentioned, this may bring all the measurement more consistent. Yes, it is!

Comparison between Hao's measurement and Quick Reduce:

Angelo sent me the fwhm, e1, e2 and R50 from quick reduce catalog based on the stars listed on this page. I compare the fwhm, ellipticity and R50 between my sextractor run and Quick reduce run. They show very similar statistical properties, though individual ones may different. Then I use my run of sextractor result to compare with other measurements statistically. Due to the great similarity between my run of sextractor and the Quick Reduce, the results on
can be directly applied to the QR results.

  • exposure 154431

  • exposure 154442

  • exposure 154449

  • exposure 155268

  • exposure 155309

Scaling between fwhm and R50 based on sextractor measurement:

Based on the runs on 11/22/2012, I made the scaling relation between fwhm and R50 based on the sextractor run. The results are show in the following plot. Based on this, the 0.9 arcsec fwhm is translated to 0.522 arcsec in R50.