Draw centering info on individual camera images

[mgalloy]

Use 20210703 as test data.

Tasks

For quicklooks:

• do not 'fill' the occulter when you draw the circle at the occulter
• change circle drawn around occulter to green
• the '+' sign in the center of occulter is using the array center, but should use the occulter center

For level 2 UNMASKED GIFS (produced about every 20 minutes):

• draw a circle at the height of the occulter in the color green (do not use the /FILL option)

For level 1 unmasked GIFs (produced about every 20 minutes):

• create unmasked GIF per camera with centering annotations at the same cadence as the other unmasked GIFs

Questions

• This is done on the realtime quicklooks in the level0/quicklooks directory. Is this sufficient or should there be an option to produce annotated level 1 or 2 quicklooks? No, need specific changes from @jburkepile.

[jburkepile]

The level0 quicklooks have a circle drawn at the location of the photosphere based on the center found for the occulter.

• Please add another circle to locate the outer edge of the occulter; you may need to make that circle red.
• The level 0 quicklooks have a green '+' at a fixed position (array center?). This is most obvious when you toggle the first and last quicklook gif of the day. Can you put the '+' at the occulter center?
• Please draw a circle at the height of the occulter in the Level 2 nomasked gifs. You create an nomask level 2 gif about every 15 minutes.

[mgalloy]

@bberkeyU, below are examples of the gallery and normal quicklooks. I am going to get rid of the grey mask under the occulter for the normal quicklooks — do you want that for the gallery quicklooks also?

Gallery quicklook:

Normal level 0 quicklook:

[mgalloy]

@bberkeyU is OK with the same as normal quicklooks.

[mgalloy]

Here are the updated level 0 quicklooks:

[jburkepile]

Thanks for making these. The data are still masked. You are applying a mask in the quicklook code so it is tough to determine if the green circle is located at the occulter edge.

To verify the level 1 pipeline centers we could save unmasked 'level1' gifs from each camera with a circle drawn at the occulter, before you shift, rotate and combine the two cameras. That would provide the center found for the calibrated images. We can talk about that to make sure its worth it.
Thanks again!

[mgalloy]

Below are the quicklooks without masking. Note that not masking changes the display scaling of the GIF because we take the display maximum from the maximum value in the array. Should we scale according to the max value in the masked version?

[jburkepile]

Wow Mike! That was incredibly fast. The unmasked are much better for assessing the accuracy of the occulter size found by the code. I like your idea to use the masked scaling on the unmasked gif. That's the best of both worlds.

Thanks so much!

[mgalloy]

Here are the new unmasked quicklooks scaled from the masked pB.

[mgalloy]

Here is the updated nomask GIF:

[mgalloy]

Full results are in raw.latest/20221007.

[detoma]

I am not sure you can use level 2 polarization data to estimate the goodness of the radius. There is too much done to the level2 data. I would go back to Stoke I and do it for the two cameras separately.

[jburkepile]

Giuliana: You make a good point. Mike measures the radius in the level 1 pB (after distortion, etc. ) Mike also saves the level 1 intensity image. He can measure the radius of the intensity mage along with the pB image for comparison Thanks very much for pointing that out.

[mgalloy]

I have a test run in raw.latest for 20240409 with the new level 1 nomask GIFs.

[mgalloy]

I was sending the same image into each GIF creation, so getting camera 0 image for both level 1 nomask GIFs.

Also, need to put logic in for making the level 1 nomask GIFs at the same cadence as the regular nomask GIFs, not for every level 1 file.

[mgalloy]

The first nomask image (for each camera) is showing green in the saturated areas and annotations, as well as the occulter circle.

• Only the first camera 0 and camera 1 nomask images have the extra green.
• The color table in the GIF is correct. Only color 255 is green.
• The value in the saturated areas of the GIF is 255.
• The annotations appear green as well.
• When I run the code outside of the pipeline, no extra green is produced.

[jburkepile]

Mike. Please create cam0 and cam1 unmasked images (FITS/GIFs) with circles in the GIFs for these dates:

• June 18, 2015
• June 22 2015
• June 23, 2015
• Oct 30, 2017
• Nov 1, 2017
• Nov 2, 2017
• Dec 16, 2019
• Dec 17, 2019

I need to look at the individual cameras with the circles for these dates to assess the changes in measured occulter radii that may indicate a platescale change. I will note this in ticket #367 where we are studying that issue.

[mgalloy]

I forgot to mention that the test dates above are done, and the results are in raw.draw_circle. The green color on saturated values is fixed, so this issue can be closed if those test dates are good.