Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-15178

Investigate position errors in scarlet blends

    Details

    • Type: Story
    • Status: Done
    • Resolution: Done
    • Fix Version/s: None
    • Component/s: meas_deblender
    • Labels:
      None
    • Templates:
    • Story Points:
      5
    • Sprint:
      DRP F18-2
    • Team:
      Data Release Production

      Description

      Lorena Mezini, an undergraduate working with Erin Sheldon, noticed that when scarlet fits positions, the mean position errors are ~0.15 in both x and y (see attached plot). What's stranger is that one of the objects in her simulations seems to have a mean positional error of 0 in x and y while the other shows the offset.

      Her dataset were generated using galsim, with very low noise and fixed bounding boxes (25 pixels) on single band images. The sources both have very small ellipticity (~.01) with random orientations, the same shear, and identical FWHM. The plot shows an example with 100k blends, each with 2 sources separated by 11 pixels.

      This ticket is to investigate this behavior to understand the conditions that lead to two seemingly identical objects being deblended differently.

        Attachments

          Activity

            People

            • Assignee:
              fred3m Fred Moolekamp
              Reporter:
              fred3m Fred Moolekamp
              Reviewers:
              Peter Melchior
              Watchers:
              Fred Moolekamp, Peter Melchior
            • Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved:

                Summary Panel