So this is "discovery" number two. It's pretty neat to see something moving in the images and have no idea what it is until you investigate. Juno is the 11th largest asteroid in the solar system and the third asteroid ever discovered (hence its 3 designation from the MPC). Unlike Mars which moves somewhat obviously against the background of stars from night to night, you would never see Juno without some help - I wouldn't at least. It's even a little challenging to see it on the aligned frames in the second image set. My SkyDelta code picks it up though (along with some noise).
1. Cropped and processed All Sky images. 2019-02-03 through 2019-02-06 at approximately 8pm local time.
2. Image set 1 frames aligned. You can see the motion here if you look close enough.
3. Raw SkyDelta motion detection via stationary frame subtraction from aligned frames [image set 2]. Juno is the only linear sequential feature. The rest is noise.
4. Linked object motion detection. This is Juno.
5. Here's one of the original frames marked up with some references on it. Juno is marked by the arrow.
6. And here's Stellarium's sky configuration for the same evening with the same references. Juno actually looks a little too close to the circled refrences compared to the actual observation, but it is by far the brightest moving object in the sky anywhere in this vicinity so I'm pretty confident it's Juno we're looking at.
- Basically the same process as Mars. I used a smaller point size median filter (2px Juno vs 3px Mars) to despeckle.
- I've seen some bad pixels in other images despite attempts to process them out. Sometimes these bad pixels look like motion when you align the frames. However, if you look back at the original frames [image set 1] you would see a bad pixel as a fixed (stationary) point. If you can locate Juno in the original unaligned frames and put your mouse cursor over that point in the image you'll see that it's definitely moving from frame to frame which rules out artificial motion from bad pixels.
Make SkyDelta detection link up closely spaced sources in sequential frames [image set 3]. I coded up a naive spatiotemporal linker and redid image set 4 to illustrate the Juno detection. The linker basically looks for sequential frames with blobs that are within some pixel distance of blobs in the previous frame. There is much more work to be done here, but it filters all the noise for these observations.
- Is Stellarium's orbit data stale or is the discrepancy in image set 6 from something else? Or am I dead wrong and this isn't even Juno? I need to eventually be able to solve these star fields to get coordinates out of them, but the wide field view plus fish-eye nature of the camera make this difficult. Astrometry.net, for instance, never solves.
- Stellarium shows Juno with an apparent magnitude of 8.9 for this observation date. If that's correct, there should be more bright asteroids that I could find with this camera. That's exciting.
Published: April 13, 2019; Updated: April 27, 2019