Friday, June 14, 2019

#189 - Tuesday, June 11, 2019 - Jupiter Double-Shadow Transit

I read in Sky & Telescope a couple months back that June 11th was going to host a special astronomical event: a double-shadow transit on Jupiter!  This happens when Jupiter's moons pass in front of it when Jupiter, Earth, and the sun are position together in such a way that we can see the shadows over the cloudtops.

Borrowed from Sky & Telescope

These occur as often as once or twice a month, but this was my first time catching one!  

Now, unfortunately, Jupiter's prime time at opposition (when the sun, Earth, and Jupiter line up, so Jupiter as at about its closest point to us) this year also comes when the plane of the ecliptic (the plane on which the solar system planets reside) sits quite low in the sky.  Jupiter only gets as high as about 26 degrees from middle latitudes, which means we have to try and view it through the denser, mushier part of the atmosphere.

The sky was relatively clear that night, with some well-formed but thin clouds drifting through, so I got set up in my front yard at around 10:15 PM.  This included my Celestron 8-inch Schmidt-Cassegrain on my Celestron NexStar SE mount, as well as my ZWO ASI1600MM Pro camera, and my Microsoft Surface 3 tablet to run the show.  I was ready to roll at 10:30.  Jupiter was still behind a tree, so I started with the moon for a few minutes.  Then I went ahead and slewed over to Jupiter to get set up, get the parameters figured out that I wanted, and practice a couple times doing the filter change, target name change, exposure time change, and capture data rotating sequence I would need to do quickly in order to get through all four LRGB filters before Jupiter rotated too much.  I cropped the frame down to 640x480 so that I could get as much speed as possible, which also allowed me to capture more frames in the moments of clear seeing.  I averaged about 20-60 fps depending on which filter (and thus, which exposure time) I was using, thanks to the USB 3.0 on my ZWO.  

Now, of course, all of those clouds I mentioned were in the south, where I needed to see!  Fortunately, they drifted north, and I finally got a clearer view of Jupiter.  I took LRGB datasets about every 5-10 minutes for an hour and a half.  I couldn't see Io's shadow in the camera image, but I could see Ganymede's!  

All told, I collected 14 datasets.  At the end, I took the camera off and put an eyepiece on, since I hadn't actually seen a shadow transit at all before.  Io was just off the edge of the disk, having completed its transit, and I could clearly see Ganymede's shadow.  It was really neat.  I frst looked through a 25mm eyepiece, then 13mm, and then 13mm with a Barlow.  The seeing was pretty bad, but sometimes I could see  more detail.  It was a fun event!  And I was quite tired at work the next day!  The temperature was quite pleasant though - 63 degrees, and no wind at all.

I'm still plowing through all of those datasets, hoping to find one that had enough clear moments to make a nicer image, but I got some not-too-terrible ones!

Frames: L: 475/2013
    R: 371/1006
G: 398/1002
B: 470/1005
Exposure: L: 12 ms
  R: 30 ms
  G: 30 ms
  B: 40 ms
ISO/Gain: 139
Stacking program: RegiStax 6
Processing program: PixInsight 1.8.6

You can see Ganymede's shadow up near the top, and I think I see Io and its shadow in the upper dark red band, closer to the right.

So the way that I processed this is:
- Stacked and did wavelet deconvolution on each LRGB channel in RegiStax
- Imported those TIFFs into PixInsight and aligned them using the script FFTRegistration that I found while I was poking around for any scripts or processes that could align on objects instead of stars.  This script can align objects as opposed to stars using a fast Fourier transform algorithm that looks at where the spatial frequencies are.  I actually did this in Matlab for a homework assignment for one of my classes once, and it was really neat how it worked!
- Applied the LinearFit process to each channel with one channel as the reference to make them all the same brightness
- The images come out of RegiStax as RGB, even though they appear grayscale, so I had to use the ConverToGrayscale process to convert them to be actually grayscale (single-channel) before I could combine them
- Combined them with LRGBCombination, but then decided to leave off the L channel, since it made it fuzzier
- Sharpened the image with MultiscaleLinearTransform
- Denoised with MultiscaleLinearTransform
- Denoised with ACDNR
- Adjusted curves and saturation with CurvesTransformation
- Calibrated the color with ColorCalibration, using the whole image as reference, and a patch of black background as the background.  This actually did a nice job, even though I think it's normally meant for stars and galaxies.

And voila!  Pretty neat.  This was my first time post-processing planetary images in PixInsight, and I liked it.  Far easier than trying to align each channel myself in Photoshop!

I'm also planning on putting all of the images together into a video to show the transit and Jupiter's rotation.  It might be a while before I can get to that, so keep your eyes out!

[ Update June 20, 2019 ] 

Finished generating the rest of the images!  Here's one of my faves.

In this one, you can easily see both Ganymede and its shadow near the top, and Io has just cleared Jupiter's limb on the right side.  

I also created a video!


So cool!!!
I made the video in PixInsight using the Blink tool.  It lets you use ffmpeg to generate video from the individual frames, and you can set the frame rate.  It does a variety of formats, so I made an AVI and a GIF.  The GIF makes it a bit noisier, but it's easier to share.

I tried to align the rotation (it appears to rotate as it moves across the sky because I used an alt-az mount), but turning on the rotation alignment in FFTRegistration didn't seem to do anything  I could rotate by hand, but that does not sound fun.  I'll find another program out there.  Or, I do have some Matlab code from a class I took where we used the Fourier transform in log-polar coordinates to calculate the rotation...

Hopefully the next one I can catch will have better atmosphere!





No comments:

Post a Comment