[KIDS] completeness/contamination
G. Sikkema
sikkema at astro.rug.nl
Wed Feb 6 11:36:50 CET 2013
Hi Francesco,
thanks for producing the new plots and clarifying the meaning of the plots,
I tested the new satellite code on this r-band image: it seems that the
satellite track is now detected and removed. We currently reprocessing
this field (as well as the g and u band). This particular satellite is
relatively narrow and bright. We are wondering how such plots look if a
faint (narrow or broad) non-detected satellite track is visible in the
fov.
We should discuss the satellite detection issue probably this afternoon
at the telecon.
And of course also the optimum set of parameters MIN_AREA and
DETECT_THRESH which are directly related to completeness and spurious
objects.
GS
On 02/06/2013 10:12 AM, Francesco La Barbera wrote:
> Hi Gert,
>
> On Tue, Feb 5, 2013 at 4:54 PM, G. Sikkema <sikkema at astro.rug.nl
> <mailto:sikkema at astro.rug.nl>> wrote:
>
> hi Francesco,
>
> thanks for the analysis report on completeness for field
> KIDS_135_-1.5.
> We are currently reprocessing the g,r,i tiles, to see if we can
> remove the satellite tracks.
>
> However, the new satellite code has problems removing very faint
> satellite tracks (at ~2 times the nosie).
>
>
> OK, thanks for the update.
>
> So in my opinion, it would be good to set the detection threshold
> at least to 2 and min_area at 5.
> Could you also send the source distribution plots (i.e.:
> CONTAMINATION_xy_min3_thre2.0.jpg ) for the following configurations?:
> min_area=5 & detect_thresh=2.5
> min_area=5 & detect_thresh=2.0
> min_area=3 & detect_thresh=2.5
>
> And from what band was your source distrubtion plot derived?
>
>
> Pls, find attached here the requested plots. I included all four cases
> (min_area=3 and threshold=2.0), as I redid
> the plots with a different color coding:
> black -> all contaminants in the selected region
> red -> contaminants with S-Ex FLAGS=0
> blue -> contaminants with S-Ex FLAGS between 1 and 3 (i.e. deblended
> sources, and/or objects contaminated by close neighbors)
>
> Also, all these plots refer to r band, as reported in the new filenames.
>
> Cheers,
> Francesco
>
>
> GS
>
>
>
>
> On 02/05/2013 12:51 PM, Francesco La Barbera wrote:
>> Dear All,
>>
>> we performed a more detailed study of completeness/contamination
>> of our KIDS catalogs by comparing
>> KIDS vs. CFHT photometry in the field of KIDS_135_-1.5.
>>
>> Here it's the steps we followed:
>>
>> 1) We downloaded the last version of the images, that shouldn't
>> be affected
>> by the satellite track issue we found in the previous version of
>> the coadd's.
>> 1) We ran KIDSCAT on the r-band images, with 6 different setups,
>> changing DETECTION_MINAREA
>> and DETECTION_THRESHOLD in the S-Ex .configuration file. We
>> considered:
>> MINAREA: 3, 5
>> THRESHOLD: 1.5, 2.0, 2.5
>> Besides that, we also ran KIDSCAT on the g and i band images with
>> MINAREA=3, and THRESHOLD=2.0, 2.5
>>
>>
>> 2) The catalogs were matched to the CFHT ones (pls ask Mario for
>> details)
>>
>> 3) The first attached plot, CFHT_KIDS_....._THRESHOLD.jpg,
>> summarizes the completeness/contamination with different setups
>> for the r-band coadd.
>> Top and bottom panels correspond to MINAREA=3 and 5, respectively.
>> For each panel, the solid curves plot the fraction of spurious
>> sources (i.e. objects with no matching to CFHT)
>> as a function of magnitude. We considered magnitudes in the same
>> aperture of 5'' in both catalogs (KIDS and CFHT).
>> Curves with different colors correspond to different
>> DETECTION_THRESHOLDs as labeled.
>> The completeness mags, for each THRESHOLD value (i.e. color), are
>> marked with a vertical dashed line.
>> (Here completeness is defined as the mag value where CFHT counts
>> keep increasing, while KIDS counts remain constant/decrease).
>>
>> 4) The second attached plot shows the spatial distribution of
>> spurious sources in the area "previously" affected by the
>> satellite track.
>>
>> 5) We would conclude as follows:
>> A) for THRESHOLD=1.5 contamination increases significantly, up
>> to ~20-22% around the completeness mag
>> B) completeness improves by ~0.15-0.2mag when moving from
>> THRESHOLD 2.5 to 2.0, and by the same amount from 2.0 to 1.5.
>> C) changing MINAREA to 5 we get slightly lower completeness
>> (by~0.1-0.15mag), and slightly lower contamination (by a few percent)
>> D) the satellite track issue is still there
>>
>> Depending on what is the maximum level of contamination we can
>> allow in the catalogs, we think we might adopt either
>> THRESHOLD=2.5 (current value)
>> or THRESHOLD=2.0, as in the latter value we get slightly better
>> completeness, at the price of a bit worst contamination.
>> Notice that the contamination fractions are indeed upper limits
>> (see our previous email on the subject), as there are many sources
>> not present in the CFHT catalogs because of (e.g.) deblending
>> issues.
>>
>> Finally, the last attached plot,
>> CFHT_KIDS_COMPLETENESS_CONTAMINATION_THRESHOLD_g_i_bands.jpg,
>> shows completeness and contamination
>> for g band (upper panel) and i band (lower panel), both for
>> MINAREA=3. Resuts are similar to those for r band.
>>
>> Eagerly waiting for your feedback, so that we can finalize the
>> S-Ex configuration file.
>>
>> Cheers,
>> Francesco, Nicola, Mario
>>
>> -------------------------------------------------------------------------------
>>
>>
>>
>> _______________________________________________
>> KiDS mailing list
>> KiDS at astro-wise.org <mailto:KiDS at astro-wise.org>
>> http://listman.astro-wise.org/mailman/listinfo/kids
>
>
> --
> dr. Gert Sikkema
>
> Kapteyn Institute
> Zernike Building, room 133
> P.O. Box 800
> 9700 AV Groningen
> sikkema at astro.rug.nl <mailto:sikkema at astro.rug.nl>
> homepage:http://www.astro.rug.nl/~sikkema <http://www.astro.rug.nl/%7Esikkema>
> 31-50-3634548
> MSN: add:cubensi at hotmail.com <mailto:cubensi at hotmail.com> to your MSN
>
> * Popular lectures on:
> - Einstein's theory of Relativity
> - Globular Clusters
>
>
> _______________________________________________
> KiDS mailing list
> KiDS at astro-wise.org <mailto:KiDS at astro-wise.org>
> http://listman.astro-wise.org/mailman/listinfo/kids
>
>
>
>
> _______________________________________________
> KiDS mailing list
> KiDS at astro-wise.org
> http://listman.astro-wise.org/mailman/listinfo/kids
--
dr. Gert Sikkema
Kapteyn Institute
Zernike Building, room 133
P.O. Box 800
9700 AV Groningen
sikkema at astro.rug.nl
homepage: http://www.astro.rug.nl/~sikkema
31-50-3634548
MSN: add: cubensi at hotmail.com to your MSN
* Popular lectures on:
- Einstein's theory of Relativity
- Globular Clusters
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://listman.astro-wise.org/pipermail/kids/attachments/20130206/0443c8c0/attachment.html
More information about the KiDS
mailing list