[KIDS] Fwd: star/galaxy separation

Francesco La Barbera flabarber at gmail.com
Mon Nov 26 10:59:01 CET 2012


Dear All,

just some update on progresses with KIDSCAT .

I just finished to implement star/galaxy separation (S/G), and I'm
rerunning it over all the 52 catalogs.
Just to remind you that the S/G algorithm is meant to be a simple one, to
obtain a "reasonable" (i.e. general purpose)
separation of the two classes.

The attached plot shows an example of how S/G works. The plot shows
CLASS_STAR from S-Ex (second-run cartalog)
as a function of MAG_AUTO S/N ratio.
In the S/N range where sure stars are available (red dots in the plot), the
code applies the biweight estimator to define the
CLASS STAR location, theta, and its width, sigma. Then we define a lower
envelope as theta-4sigma. If S/N is below that of
sure stars, we compute a running median trend, including only objects with
CLASS_STAR>0.8, and shift it downwards to match
the theta-4sigma locus.
The resulting blue curve in the plot defines the separation of stars and
galaxies (i.e. objects with CLASS_STAR > and < the blue curve
are tagged as stars and galaxies, respectively).

The blue curve is stored in an output file, OUTPUT_RES/SG_CURVE.dat (ascii
format).
The attached plot is automatically produced for each coadd (PLOTS/SG.ps).
The 2DPHOT FLAG in the catalog is updated accordingly. If an object is
classified as a star, the 2DPHOT flag value is 4. Different
flag values are summed up (e.g. 2DPHOT_FLAG=5 means the object is a sure
star, and is also classified as star according to
the above procedure).

>
Pls, lemme know whatever feedback/comments you have, so that I can change
program/scripts accordingly. Also, if you like,
I can share the (fortran) code.

>
The catalogs of stars and galaxies will then be used to define
completeness...........

>
Cheers,
Francesco
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://listman.astro-wise.org/pipermail/kids/attachments/20121126/b928686c/attachment-0001.html 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: SG.ps
Type: application/postscript
Size: 561810 bytes
Desc: not available
Url : http://listman.astro-wise.org/pipermail/kids/attachments/20121126/b928686c/attachment-0001.ps 


More information about the KiDS mailing list