Minutes: 16th March 2006
My apologies for the late appearance of these minutes. Their fidelity
may have been significantly degraded by the delay.
Present: RHL, Steve Asztalos, Dave Monet, Tim Axelrod, Kem Cook, Ed
Ol.,
David Burke, David Bazell, Nicole Silvestri, Gary Bernstein, Andy Becker
Overall Topic: Shape Measurements
PSFs:
Gary doesn't think that the Bernstein/Jarvis PSF model (i.e. modelling
the variation of the PSF at a given point in the focal plane as a
function
of time/seeing) is the last word; could be merged with an approach
to spatial variability such as the SDSS K-L modelling.
RHL: what can we learn about PSF modelling from the simulations?
David Burke (?): use the SLAC Gemini data [did he mean in conjunction
with sims?]
Becker: we're not there yet
Tim: sarcastic unreasonably comment about RHL's dereliction of duty in
helping Phil Pinto use Chuck Claever's ray tracing to model the PSF.
RHL: who's working on this?
All: U. Washington; Bernstein and Jarvis; RHL; Steve Asztalos and
David Burke
Shape Measurements
RHL: For LSST there are two branches that I can think of:
1/ Measure some moments for shape/polarisability (son-of-Kaiser?)
2/ Fit a model to the above-the-atmosphere image and estimate
that model's properties
Gary: I'm using the most expensive version; fitting n individual images
using Gauss-Laguerre basis functions for the galaxy and the PSF [this
means that the PSF convolutions are fast]. This takes c. 1second/galaxy
for a 20-image DLS stack on a GHz-class Pentium
Tim: not a problem
RHL: How well will these models work for galaxy photometry? I have
an unwritten paper on how well v. crude models reproduce Petrosian
photometry for even bright galaxies in SDSS.
Gary: it should work well
Tim: Chris Roat at UCSD's about to run a similar algorithm on DLS
using Gaussians; we should ask him about it in a month
[RHL note added in proof: that's only a week from the appearance
of these minutes]
RHL: how about using models for morphological measures? The classical
anisotropy etc. are aperture measures and are noisy. If I wanted to
play with this, how hard would it be to re-implement your work from
scratch?
Gary: There are Tricks... ask me for the code.
What Is Next?
RHL: That wraps up the discuss-a-topic part of this 'phone con's life
cycle.
I'd like to move on to focussed discussions of progress... starting with
Andy Becker on Image Subtraction with HOTPANTS. This will be on the
6th of April.
_______________________________________________
LSST-data mailing list
[email protected]
http://www.lsstmail.org/mailman/listinfo/lsst-data
