Details about Callisto :
The transformers make it possible to image in any type of weather, here atmospheric turbulence and fog.
Tomorrow, the Jetson Orin Nano will allow this to be carried on board :
1 Like
Seems too beautiful to be true. Canât get such result with significant turbulence.
1 Like
Itâs normal, your technique based on âlucky imagingâ depends on how lucky you are to capture good pieces of planet.
Here, I no longer consider luck but the ability of âtransformersâ to deal with distortions, blurring and noise.
When the Jetson Orin Nano 8GB card is available, I will be able to make the following animation using its AI computing power.
1 Like
We have to be serious a try to avoid exchanges like the ones you had on webastro website.
You can get some improvements with quite good RAW captures but you canât reconstruct good and detailed images with high turbulence blurry images. Only adaptation optic can do that.
I donât say your work is bad. I just say you have to be more precise with the information you give and try to do not sell dream.
I guess i will stop here and not start an other webastro.
Alain
1 Like
Thank you for your wise advice. We are a small group working on atmospheric turbulence mitigation.
I understand very well the bitterness of some people, including you. A site will soon be online to show the progress of the work and in the meantime I ask you not to intervene on my publications. Thanks in advance.
1 Like
Another version of image n°112 obtained by an amateur astronomer using the same technique (transformers).
The planetâs poles and limbs are better defined than with the traditional method using lucky imaging.
Jupiter October 18 in the fog. The degradation model integrates fog processing in addition to atmospheric turbulence.
Artificial intelligence through transformers makes it possible to deal with different degradations of the assisted video.
The Io satellite crosses the planet.
End of Ioâs transit in front of Jupiter, October 19, 2022 at 00:18:50.
A new processing containing some improvements: finer alignment of adjacent frames and taking into account fog in the degradation model.
Mars this morning :
1 Like
A softer version :
Jupiter and Mars on November 8, 2022 between two cloudy periods :
C11@f/20 + ZWO ADC + ASI 462MC + EQ-R Pro + turbu v0.6 software.
Jupiter of November 10, 2022 step by step:
-
accumulate 6000 raw images
-
reveal the details of the previous image
-
search for new details in depth (8+1+8)
-
remove residual blur and strengthen the final image
⊠all this live and now the video
My new acquisition technique no longer uses lucky imaging but (8+1+8) x 6000 = 102000 raw images to create one final image and 86 x 102000 = 8772000 raw images for this video.
Hello.
Your work is very interesting. Could you please provide your raw images to compare with conventional treatment.
Sincerely
Hello,
No raw images stored but the accumulation of 6000 raw images.
Below are the ZIP files :
jup1.zip (99.4 MB)
jup2.zip (97.4 MB)
Iâm curious to see what another method does.
Regards.
Thank you for your reply.
Unfortunately, it is not possible to test with these files. Working with files already processed by your application will not show anything.
Additionally, images sature in the green curve and are at the limits of the blue curve.
As a result, attempts of treatment will necessarily be a failure.
For each saturate pixels, information is lost.
Regards,
You need to find a way to reveal the details hidden in the previous blurry images, this is the second step:
sharp.zip (46.0 MB)
Opposition 2022 of Mars, 5 hours video :
1 Like
Mars as if you were there, with very strong turbulence conditions and a 40m/s jetstream.
These 5 videos represent my computer screen. I regularly use a Geforce GTX 1050 Ti card but it also works with the Jetson family including Nano 4GB.
My Dropbox :