Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-38632

Create task to measure Normalized Compensated Gaussian flux



    • Story
    • Status: In Progress
    • Resolution: Unresolved
    • None
    • meas_algorithms
    • None
    • Data Release Production
    • No


      When DM-32994 is done we will have a new measurement plugin that can measure a Compensated Gaussian Flux, which has high signal-to-noise on stars and is robust to background offsets. However, experimentation and sims have shown that this can vary relative to the "total" flux (as estimated from either large apertures or psf models) not only in overall normalization (which is fine and arbitrary) but also with some spatial structure over the typical scale of an HSC or LSSTCam detector. (Testing with LATISS is TBD).

      The plan is to write a task in this ticket that (a) measures the compensated gaussian (CG) flux; (b) measures a large aperture flux; (c) estimates the aperture correction to put the CG flux on the same scale as the large aperture flux; (d) applies this aperture correction to output a "normalized compensated gaussian flux" that is suitable to be used for all calibration in the new CalibrateImageTask being stood up in DM-38546. (The expectation is that this new task in this ticket will probably not be implemented using the existing characterize/calibrate tasks, but only in the new task).

      Note that this plan does have a shortcoming: in particular, the large apertures that will be used to define the aperture correction/normalization are susceptible to background offsets. As much care will be taken to ensure these problems are minimized (as in DM-37955, but see also PREOBS-3327 and DM-37912). Furthermore, it is my expectation that the photometric calibration (local or global) will be able to handle any small offsets that arise because of these small background offsets in the normalization aperture correction because (a) all calibration is relative anyway; and (b) applying psf/cmodel/gaap/etc aperture corrections relative to the NCG fluxes will be much more reliable because they will not be affected by background offsets. However, much of this will have to be demonstrated after this task is written.


        Issue Links


            There are no comments yet on this issue.


              erykoff Eli Rykoff
              erykoff Eli Rykoff
              Eli Rykoff, Jim Bosch, John Parejko
              0 Vote for this issue
              3 Start watching this issue




                  No builds found.