2016_09_16 visualization meeting


(Jon Thaler) #1

Bluejeans: 107541143

Agenda:

  • UIUC status.

Current responsibilities:

  • @weiren2:

  • Infrastructure and Front end - Back end communication

  • Modify JSON file to accommodate gain and offset scaling by segment.

  • Remove all hard-wired quantities. Standardize FE/BE communication.

  • Tomcat server for the new API. We already have two Tomcat instantiations.

  • @ywang

    • Image math (sum, difference, ratio) with display of result.
    • Back end histogram generation. Second moments of CCD amplifiers will be the first demonstration.
  • @jbpagliuco:

  • FE: histogram display. Is partially implemented. Second moment will be the demo.
    Add functionality: Rescale x and/or y axes. The first requires communication with the back end.

  • Commands driven by cursor location. Partially implemented.

  • Interface to the new Firefly API. We’ll use a new branch.

  • We need a small image repository for testing.

Notes:

  • We need a repository of raft-level images. @tony_johnson is assembling some from CCD images taken at BNL. We want to be able display raft-level coordinates. The FITS metadata will include the required information.

  • There are lots of open Github issues. Some are quite old. Are they all really unfinished? We need to clean them up.

  • Gain/offset corrections will be done by Tony’s FITS generator. He will put the parameters in the metadata.

  • We need to have two versions of images: with and w/o overscan.

  • Firefly has a histogram functionality.
    Action: @jbpagliuco: Look at it, for discussion next week.
    Action: @jjt: Make a list of use cases, to assess its applicability.

  • Next week: DIscuss result of Xiuqin’s meeting with DM folks (Jacek, Gregory, et al.).


2016_09_23 visualization meeting
(Jon Thaler) #2

Here is my first pass at a list of histogram use cases

Jon

Many will require region specification.

  • Pixel distribution.

  • In an area or line.

  • Spatial distributions of pixel values
    Perhaps with baseline (eg, overscan or sky) subtraction.

  • Radially from a point.

  • y (or x) projection in a 2D region.

  • Distributions of analysis results:

  • Noise by amplifier in a CCD.

  • hot/dead pixels by amp in a CCD.

  • History plots (eg, noise vs time in each amp)

  • Comparisons with templates or other images:

  • Ratio of present CCD amplifier noise to previous values.

  • Average pixel in an amplifier vs image illumination (for gain
    measurement)

  • Pixel dependence on other metadata quantities

Features that may apply to many (or all) use cases:

  • x and/or y scaling (zoom to an interesting part; ignore large bins)
  • Overlay of multiple histograms
  • Overlay of function on a historam (eg, results of fit to the
    distribution)
  • Axis labels
  • We will probably want to be able to choose between a linear and
    logarithmic scale for the y-axis.

Priority:
1: Pixel dis1. tribution in a region.
2. Analysis results (eg, noise vs amp).
3. x or y projection (eg, y distribution in an overscan region)
4. Comparison with a template (subtraction, ratio, or overlay).