Stacking centered on comet nucleus

[abillionnewstars]

Since it is comet time, a live stack aligned on comet nuclei would be amazing to have!!

And while you are at it, an option to remove the stars as the stacking is live.

Easy peasy!! Best ideas come from the -18°C night chasing a comet

Cheers

Stephan

[admin]

Hi Stephan,

if your tracking of the comet is spot on then you could just turn off alignment in the live stacking settings and let the stack build up. In theory the comet should stay still and the stars should trail out. If you save the raw frames too then you can feed them into DSS later for a more thorough alignment.

I guess a 'real' comet stacking mode would require some way to detect the comet position in each frame. Perhaps if we could eliminate the stars and the background then look for the center of brightness of what is left? So far I haven't tried for the comet (last time I checked it was going to rise so late that it was actually early, and I rarely do that!) If you or anyone else has some unprocessed subs that they would like to share then I can have a play to see how easy/hard it is to detect the comet.

cheers,

Robin

[abillionnewstars]

Hi Robin,
That's interesting.
I certainly can share subs. They are 30 seconds long. I imaged it for an hour. How many would you like to have? In what format?

Cheers

Stephan

[Jean-Francois]

Hello Stephan and Robin,

I can propose a different method with a lot of calculations ... nice for the people that love to code something

It is possible to calculate for each comet the direction vector (arcsec per sec in RA/DEC) in the sky (note the vector change with the time).
This displacement can be calculated in pixel X and Yon the image after a plate solve of the image.
The image can be rotated or mirrored ... the stacking program can finally shift the individual image (after a quick intermediate plate solve)
with the opposite direction. The stacked image will show the stars moving while the comet will stay in place.

If the live stack in short, then the mount can have an auto-guiding on a star.
If the comet displacement is quick, then the comet will move outside the field of view of the camera.
Note that some mount drivers allow to set a RA and DEC rate offset, so that the mount "follows" the movement of the comet.

Regards,
Jean-Francois

[admin]

Hi Jean-Francois,

interesting approach - basically offset each frame from a normal stack based on the known RA/Dec drift rate of the comment.

@Stephan,

Five frames in FITS format from different points in the run (say 10 minutes apart) would be enough to have a first look.

thanks,

Robin

[admin]

This is what we are looking for, yes...?

cometalign.JPG (90.73 KiB) Viewed 4824 times

Robin

[admin]

Right, so since the comet is active right now, I am trying to get an updated build together - if it works first time, I will upload it this evening, otherwise it will have to wait for tomorrow...

How it works...

Select live stacking as usual, make sure alignment is turned on and check the 'Align on a Comet' box. This gives you two new options to adjust

* Star removal filter size - set this to a large value - 50+? Any blobs smaller than this size in pixels will be removed from the image as part of the comet detection process. This will (hopefully) get rid of any stars

* Detection threshold - the brightest part of the image (this percentage of total area) will be assumed to be the comet (after the stars have been processed out). The position of the comet is calculated as the average center of brightness of this brightest portion.

As you can imagine, the detection will only work if the comet is the biggest, brightest fuzzy thing in the image.

You can tune the detection by turning on 'Highlight Detected Stars' and also the option to View the individual frames rather than the stack - that will show each frame *and* the comet detection box - aim for it to be around the nucleus, as shown above.

You will probably need plenty of memory to run this, especially on a high resolution camera - I wrote it without too much consideration for the resources being used - just trying to get something that worked quickly...

cheers,

Robin

[admin]

Here's a test version to try

Direct Link removed - please use the version on the beta downloads page : https://www.sharpcap.co.uk/sharpcap/sha ... p-4-1-beta

You should be able to stack existing subs using the folder monitor camera - that's how I was testing it.

cheers,

Robin

[abillionnewstars]

It looks very promising!! I will give it a try and get back to you!

Stephan

[Jean-Francois]

Hello Robin,

I test it ... it works
At least for a quick live stack, the function gives a good result.
But not yet for a final image. The comet position calculation may be perturbed by a background star.
The stars trail are not in a line. I compare the live stack with a stack that I did with the software Prism with the calculated image shift from the comet X/Y shift in the sky. The star lines are along a line with the result from Prism (pre-calculated image shit).
I take today only the green images from my Canon 6D from NEOWISE-2020.

SharpCap_Align_Comet.jpg (226.3 KiB) Viewed 4532 times

But after this test, I try to tune the "Star Removal Filter Size" and the "Detection Threshold" ... it is possible to find better parameters.
With "Star Removal" = 15 ... the result is much better

SharpCap_Align_Comet_2.jpg (67.88 KiB) Viewed 4532 times

Regards,
Jean-Francois