lsstdesc / cfhtls-reprocess Goto Github PK

Package : meas_extensions_shapeHSM already exists but the current version is incompatible with the latest version of the stack.

Astrometry on each image has already been implemented relative to an external catalogue. This fine tuning will act on the stack of images to find relative astrometric shifts.

Package : meas_extensions_shapeHSM already exists but incompatible with the current stack.

This require a little bit of code tuning, but should be ~under control.

We need to find a way to make the CFHT data available to our community, in the most useful/easiest way.

In order to get ready for the LSST case with a wide FOV, It is worthwile to study the sensitivity to cloud attenuation variable across the CFHT 1 deg FOV.
It implies a well mastered photometry.

We need to understand how the PSF models work.

The current definition of the PCA_Psf class in DMStack can be found here:

http://lsst-web.ncsa.illinois.edu/doxygen/x_masterDoxyDoc/classlsst_1_1meas_1_1algorithms_1_1_pca_psf.html#details

Compute magnitude and compare to CFHT catalog for the 5 filters

Here is a list of ideas and things to check by Chris Stubbs

• add sources and measure completeness and transfer functions
  photometry
  astrometry
  shapes
• compare to CFHTLS analysis, on all measured quantities
• frame subtraction and light curve tests
  find SNe and do forced photometry on them
• variable object analysis
• Precision Calibration: Look at color-airmass terms, look for color-dependent residuals in z band and compare to Gemini archival water vapor EW
• Find and catalog all asteroids (how long was expsoure time?)
• Add fake asteroids and measure recovery rate vs. frame separation
• Impose sensor artifacts and then have DM remove them
  Jim Chiang and edge rolloff
• Look for evidence of fatter-brighter effect in PSF
• Look for persistent astrometric residual vector field as evidence of lateral E-fields
• Linearity test using cloudy or bright-sky data
• compare co-added frames with single-epoch averages.
• PSF analysis in co-added frames
• Look at astrometric residuals vs. exposure time and airmass and passband
  then compare with Phosim

We want to determine how well the DM pipeline performs relative to the previous survey work.

Some ideas include comparing detection efficiency of objects as a function of magnitude compared to the truth information in overlapping Hubble fields.

the idea is to run subtraction, make some validation analyses to remove fakes, construct lightcurve, and then compare to the force photometry from the stack pipeline