Uploaded image for project: 'Data Management'
  1. Data Management
  2. DM-34364

Use PSF-matched template for image decorrelation

    XMLWordPrintable

    Details

    • Type: Story
    • Status: Done
    • Resolution: Done
    • Fix Version/s: None
    • Component/s: pipe_tasks
    • Labels:
      None

      Description

      The current ImageDifferenceTask runs decorrelation of the difference image using the warped (but not PSF-matched) template. However, the decorrelation Task expects the template to have been convolved with the PSF-matching kernel, with the result that the variance plane is too low. Fixing this bug is simple, but it should be noted that this will change the variance plane and the number of diaSources detected on the difference image.

        Attachments

          Issue Links

          relates to

          Bug - A problem which impairs or prevents the functions of the product. DM-31948 Fix regression in AP catalog metrics

          • Done

          Story - Created by Jira Software - do not edit or delete. Issue type for a user story. DM-33001 Design refactoring of ImageDifferenceTask

          • Done
          mentioned in

          Page Loading...

            Activity

            Ascending order - Click to sort in descending order
            Hide
            Permalink
            sullivan Ian Sullivan added a comment -

            PR: https://github.com/lsst/pipe_tasks/pull/662
            Jenkins: https://ci.lsst.codes/blue/organizations/jenkins/stack-os-matrix/detail/stack-os-matrix/36314/pipeline/47/

            Show
            sullivan Ian Sullivan added a comment - PR: https://github.com/lsst/pipe_tasks/pull/662 Jenkins: https://ci.lsst.codes/blue/organizations/jenkins/stack-os-matrix/detail/stack-os-matrix/36314/pipeline/47/
            Hide
            Permalink
            ebellm Eric Bellm added a comment -

            (I note that per discussion on Slack at https://lsstc.slack.com/archives/C2JPMCF5X/p1649367089692649 that we appear not be using the variance plane for detection at all, so having the variance plane wrong won't actually adjust the number of detected DIAsources. This may also explain a) why we haven't seen much effect from the variance rescaling b) perhaps some of our "no detection" artifact class.)

            Show
            ebellm Eric Bellm added a comment - (I note that per discussion on Slack at https://lsstc.slack.com/archives/C2JPMCF5X/p1649367089692649 that we appear not be using the variance plane for detection at all, so having the variance plane wrong won't actually adjust the number of detected DIAsources. This may also explain a) why we haven't seen much effect from the variance rescaling b) perhaps some of our "no detection" artifact class.)
            Hide
            Permalink
            sullivan Ian Sullivan added a comment -

            The variance plane is used for detection of diaSources in image differencing, but we also rescale the variance plane which mostly fixes this bug.

            Show
            sullivan Ian Sullivan added a comment - The variance plane is used for detection of diaSources in image differencing, but we also rescale the variance plane which mostly fixes this bug.
            Hide
            Permalink
            Parejkoj John Parejko added a comment -

            I'll trust your description in the ticket.

            Show
            Parejkoj John Parejko added a comment - I'll trust your description in the ticket.
            Hide
            Permalink
            sullivan Ian Sullivan added a comment -

            New PR in ci_hsc_gen3 to fix the failing unit tests: https://github.com/lsst/ci_hsc_gen3/pull/73

            New Jenkins run: https://ci.lsst.codes/blue/organizations/jenkins/stack-os-matrix/detail/stack-os-matrix/36397/pipeline

            Show
            sullivan Ian Sullivan added a comment - New PR in ci_hsc_gen3 to fix the failing unit tests: https://github.com/lsst/ci_hsc_gen3/pull/73 New Jenkins run: https://ci.lsst.codes/blue/organizations/jenkins/stack-os-matrix/detail/stack-os-matrix/36397/pipeline

              People

              Assignee:
              sullivan Ian Sullivan
              Reporter:
              sullivan Ian Sullivan
              Reviewers:
              John Parejko
              Watchers:
              Eric Bellm, Ian Sullivan, John Parejko, Yusra AlSayyad
              Votes:
              0 Vote for this issue
              Watchers:
              4 Start watching this issue

                Dates

                Created:
                Updated:
                Resolved:

                  Jenkins

                  No builds found.