Help with sunspot analysis (new)

SunPy
,
1

Hello I have been given an assignment to find the time period for the sun’s rotation by looking at sunspots with a telescope and was told sunpy is useful for analysing batches of images as it is faster than doing everything manually. Unfortunately, none of my peers have any experience using sunpy to provide any help and I’m struggling to apply the information on the website to what I need to do so I’m hoping the community could provide some input.

  1. Is it possible to get sunpy to find the pixel coordinates of the centre of the sun in an image?
    (this is to find celestial north, I tried using .center but it said there’s no metadata in the image so I’m not sure if I can add that.)

  2. Can sunpy find the polar coordinate of a sunspot from the celestial north previously found?
    (this seems less likely especially due to artifacts in the image so I’m not sure how it would correctly identify the sunspot unless I can do that bit manually, but it’s what I’m told I need to do)

Thank you in advance to anyone willing to provide help for such broad ended questions.

2

Hello @rkjj2022,

Thank you for reaching out. I am probably not the best placed to answer you but I will try.

sunpy can do those items but it requires that the files you load into sunpy have the correct metadata for your observation. Without that, none of the machinery within sunpy will work.

From that, I think we will need more information from about your telescope and the files it creates. From there, we can try to formulate what metadata is missing and then go from there.

How does that sound?

3

Yes that would be wonderful thank you.

I am using a Coronado Personal Solar Telescope, and taking images observing through a camera using the ICAP software, though I’m unsure what the specifics of the camera I put into the eyepiece are.

It produces TIFF files which I can view using DS9, and I have imagemagick to convert them into FITS files.

I hope this information helps, thank you for your response.

4

I’ve just recently been looking into a similar problem.

I do amateur solar imaging with my Lunt50 telescope and Saturn-M camera. After initial processing using common tools for stacking and sharpening, I’ve been working on a python script to do additional processing to enhance contrast beyond simple image-wide gamma curves and histogram equalization.

A key part of that is the need to identify the center and radius of the solar disk in the image, so that’s what I first worked on using opencv. Take a look at my script:

Look at the findValidCircle and findCircle functions, using the EDCircles algorithm. It does a good job of finding the center and radius.

Recently, I’ve been looking at SunPy and thinking how I might use it. What strikes me is that it seems very focused on data from professional instruments where all of the relevant metadata is available. For someone like me with my personal Lunt50 and you with your PST, there’s little meta data available. I was looking at this documentation for how to create a “map” from an image file:

And it seems like what I need to do is call make_fitswcs_header(), and give it info like date/time, center reference pixel, and scale of my pixels.

My circle-finding script can find the center. The scale in arcseconds can be estimated using the radius in pixels and the known size of the sun in arcseconds for the day of the image capture. Or can be determined directly if you know the telescope’s optical parameters and the camera’s pixel size. Though that might be complicated if capturing through an eyepiece.

You can use a tool to convert image file formats, but I think the important thing to understand is that your tiff files are missing important meta data that sunpy requires, so the fits files will necessarily be missing it as well. If the meta data was all there in the fits file, SunPy would automatically extract it and use it, but that’s not really an option. So instead, I think best is to piece together what info you can and feed it into make_fitswcs_header().

If I understand this bit from the example, it essentially says that the reference pixel is the center of the sun, and it’s captured from earth on a given date/time - so aside from date/time, this would be the same for every image:

coord = SkyCoord(0u.arcsec, 0u.arcsec, obstime=‘2013-10-28 08:24’,
observer=‘earth’, frame=frames.Helioprojective)

This piece then would get the specific center found and calculated pixel scale:

header = sunpy.map.make_fitswcs_header(data, coord,
reference_pixel=[0, 0]*u.pixel,
scale=[2, 2]u.arcsec/u.pixel,
telescope=‘Fake Telescope’, instrument=‘UV detector’,
wavelength=1000u.angstrom)

Probably unnecessary, but you could set telescope to the PST, instrument to your camera, and wavelength to 6562.8 angstroms for h-alpha.

Anyway, this is the sort of thing I was looking into… having a somewhat automated function to fill in this data. If there was a SunPy function that gave the size of the sun in arcseconds for a given day, that would make things simpler.

5

aha - looks like this is the function to get the angular size as a function of time:

6

OK… I had a bit of time and figured I might as well take a stab at experimenting with making a Map from an image from my Lunt50. Here’s the result:

No guarantees that it’s correct, but I think I at least got it mostly correct, to the point where I can display a grid on my solar disk and it seems to line up reasonably given my pixel size estimate.