Improved RealGalaxy catalog

[rmjarvis]

The current RealGalaxy catalog could be improved in a number of ways. (Feel free to turn this issue into some subset of these and spawn the rest off into other issues as appropriate.)

• Possible improvements to mono-chromatic RealGalaxy catalog:
  • Increase the size of the current sample.
  • Increase the depth by using deeper fields.
  • Denoise the images to make them effectively deeper while still retaining the important features of the profiles. (Note: @arunkannawadi working on this.)
• Steps to enable ChromaticRealGalaxy objects:
  • Get images of HST galaxies with reasonably consistent MultiDrizzle settings, in a range of band passes. (Note: Anton Koekemoer is going to send AEGIS data, which is a good starting point, with F606W and F814W. Later we will move to CANDELS.)
  • Make sure that CTI correction software has been run.
  • Make appropriate cutouts around galaxies that are appropriate for all bands.
  • Estimate PSFs from star fields. This requires some additional code infrastructure once we no longer use just F814W.
  • Run Sersic and B+D fits to these objects in each band, so we can have realistic parametric chromatic galaxies.
  • Assemble enough blank fields in each band to have an estimate of the CorrelatedNoise in each that we can use for whitening purposes.
  • Look into denoising the images (as above) so we can make the sample effectively deeper.

[rmandelb]

I think it's worth separating this into "work to be done with external datasets" and "work to be done in GalSim to make it accommodate these things".

I actually have work going on all of these fronts, but am happy to collaborate with others, as some of these are big things.

[rmjarvis]

Of course. Feel free to refocus this issue as appropriate and spawn new ones if need be. Just trying to get all the ideas generated in our session today into issues so they don't get lost. :)

[rmandelb]

Yes, thanks for that!

[rmjarvis]

I think #640 is now the issue for the work that needs to be done in GalSim to incorporate color information into RealGalaxy objects. So let's use this issue for the work that needs to be done to assemble the relevant information from HST images.

At the recent DESC meeting, we discussed this with @plazas, @jasondrhodes, and @jmeyers314. Andres will have a summer student who may be able to work on at least some aspect of this. I've added some more tasks to the project description at the top. Please feel free to add more if you think of some. Then Andres and Bradley (the summer student) can pick out one or two of them to work on.

[jasondrhodes]

Sounds good to me!

Sent from my iPhone

On Mar 2, 2015, at 10:42 PM, Mike Jarvis <notifications@github.commailto:notifications@github.com> wrote:

I think #640#640 is now the issue for the work that needs to be done in GalSim to incorporate color information into RealGalaxy objects. So let's use this issue for the work that needs to be done to assemble the relevant information from HST images.

At the recent DESC meeting, we discussed this with @plazashttps://github.com/plazas, @jasondrhodeshttps://github.com/jasondrhodes, and @jmeyers314https://github.com/jmeyers314. Andres will have a summer student who may be able to work on at least some aspect of this. I've added some more tasks to the project description at the top. Please feel free to add more if you think of some. Then Andres and Bradley (the summer student) can pick out one or two of them to work on.

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/551#issuecomment-76843149.

[plazas]

OK, thanks! I will look at it and coordinate with Jason and Bradley.

2015-03-02 17:46 GMT-05:00 jasondrhodes notifications@github.com:

Sounds good to me!

Sent from my iPhone

On Mar 2, 2015, at 10:42 PM, Mike Jarvis <notifications@github.com<mailto:
notifications@github.com>> wrote:

I think #640#640 is
now the issue for the work that needs to be done in GalSim to incorporate
color information into RealGalaxy objects. So let's use this issue for the
work that needs to be done to assemble the relevant information from HST
images.

At the recent DESC meeting, we discussed this with @plazas<
https://github.com/plazas>, @jasondrhodeshttps://github.com/jasondrhodes,
and @jmeyers314https://github.com/jmeyers314. Andres will have a summer
student who may be able to work on at least some aspect of this. I've added
some more tasks to the project description at the top. Please feel free to
add more if you think of some. Then Andres and Bradley (the summer student)
can pick out one or two of them to work on.

—
Reply to this email directly or view it on GitHub<
#551 (comment)

.

—
Reply to this email directly or view it on GitHub
#551 (comment)
.

[rmandelb]

@rmjarvis - thanks for expanding the project description. I just added a few more things and clarified a few others.

[rmjarvis]

I suspect any further effort in this direction will take place outside of GalSim per se with the output of some user class serving as inputs to GalSim. This isn't particularly hard to arrange in code, so I don't think we need this to be a GalSim project anymore. Feel free to reopen if there is something in GalSim proper that could help with this.