Printable Version of Topic

Posted by: Bjorn Jonsson Jul 29 2020, 12:48 AM

The images from perijove 28 are now available. This is image PJ28_24 in approximately true color/contrast and enhanced versions:

Image 28_24 was obtained shortly before Juno's perijove which makes it a very high resolution image. The resolution is highest (slightly better than 3 km/pixel) in the top center area in the 'central' image (the biggest image). At this high resolution some of the cloud features look rather fuzzy.

Posted by: Kevin Gill Jul 29 2020, 04:49 PM

First attempts at maps and composites of PJ28 imagery:

https://flic.kr/p/2jqyJDy
https://flic.kr/p/2jqyJDy

https://flic.kr/p/2jqPrP7
https://flic.kr/p/2jqPrP7

https://flic.kr/p/2jqP52C
https://flic.kr/p/2jqP52C

I've made some great strides improving my pipeline. I'm using some dark fields generated by Bjorn (Thanks!) and fixed a bug that compressed the images in an ugly way. Also, I'm getting good results integrating Denoise AI and Sharpen AI from Topaz Labs into my process. Not sure which one works the best yet, but both are really good. I'll continue uploading most of my intermediate files such as maps, meshes, Blender files, etc to my Google Drive: https://drive.google.com/drive/folders/1mWW6g6ZQomBlR5DR5pGxd53tg81F4j_L?usp=sharing

Posted by: Bjorn Jonsson Jul 29 2020, 09:46 PM

These have now become a 'standard' part of my processing pipeline as well, especially Denoise AI. Usually the results are great (much better than in e.g. Photoshop) but interestingly, I had difficulty getting results I was happy with from Denoise AI when processing image PJ28_24 (posted at the start of this thread). The problem seems to be that this image has a significant amount of relatively fuzzy cloud features plus some low contrast areas, especially where the resolution is highest near top center in the 'central' image in my post. I ended up not using Denoise for this particular image.

Posted by: Sean Jul 30 2020, 02:55 PM

My approach was to average the results from multiple upscalers/denoisers by masking and blending between them. It was labor intensive since the process had to be dialed in separately for each frame.

Posted by: palebutdot Aug 2 2020, 08:24 AM

Our amazing Gerald Eichstädt has spotted some lightning on Jupiter!
https://www.missionjuno.swri.edu/junocam/processing?id=9054

Posted by: fredk Aug 3 2020, 12:44 AM

Not a lot of detail there - can we be sure these aren't cosmic ray hits, eg?

Posted by: mcaplinger Aug 3 2020, 03:39 AM

We can't be sure, but they're generally consistent with point sources (there are four pixels with values above background), whereas cosmics tend to either one pixel or many pixels in a line.

See https://www.storm-t.iag.usp.br/pub/ACA0330/2016/Monografia/Juliana-Alvaro/galileo-jupiter.pdf especially Figure 5.

Posted by: fredk Aug 3 2020, 04:33 AM

Yeah, PSF-like does suggest a real source.

Unless I've missed this somewhere, these were imaged on PJ28? So presumably day side? I recall the nightside Galileo observations, but are lightning strikes expected to be visible on the day side?

Posted by: mcaplinger Aug 3 2020, 06:38 AM

No, night side, during the high-TDI departure imaging.

We've seen similar things in the northern hemisphere on earlier PJs, e.g., pj14-002.

Posted by: fredk Aug 4 2020, 01:31 AM

Thanks for the details. Any sense of whether these flashes are comparable in brightness to flashes seen by Galileo?

Posted by: mcaplinger Aug 4 2020, 05:37 PM

It's fairly hard to say. Galileo was taking long time exposures (tens of seconds long, IIRC) and we are taking much shorter exposures with TDI where the flash presumably only lasts as long as one pixel residence time. I guess if you assumed the flash duration was similar one could back out the relative radiometry, but I haven't done that.

Posted by: Brian Swift Aug 6 2020, 07:45 AM

Has any consideration been given to capturing long exposures with TDI disabled
(or intentionally mismatched to spin rate) to elongate or separate flashes?
I see in the GALILEO LIGHTNING paper, they did some "scanned" frames.

Posted by: mcaplinger Aug 6 2020, 03:10 PM

That's an interesting suggestion, we could look at that. Frankly, Junocam is not the best instrument on the spacecraft to be looking at lightning. We only take these images because the orbit geometry doesn't allow anything else and it's a way to keep the downlink full.

Posted by: Brian Swift Aug 6 2020, 08:51 PM

I'd like to see all the SRU images/data show up in the PDS.

Posted by: Gerald Aug 7 2020, 12:44 PM

Maybe I should provide links to the maps the gif is derived from.
http://junocam.pictures/gerald/uploads/20200730/.
http://junocam.pictures/gerald/uploads/20200731/.

Specifically, http://junocam.pictures/gerald/uploads/20200731/JNCE_2020207_28G00087_V01-raw.bmp_pol_10px_30.051500s_cx812.0_000000_taylorIncEm.bmp_polS_.html, and http://junocam.pictures/gerald/uploads/20200731/JNCE_2020207_28G00090_V01-raw.bmp_pol_10px_30.051500s_cx812.0_000000_taylorIncEm.bmp_polS_.html.
There are only very few energetic particle hits in those images. And the lightning candidates are located at FFRs, where lightnings would be plausible or expected.
So, besides the structure of the spots, the location hints towards lightning on a probabilistic basis.

In the lower third of http://junocam.pictures/gerald/uploads/20200728/, you'll find less processed drafts to look for possible energetic particle hits and for stars to compare them with lightning candidates.