Comet alignment star trail removal

[blackdragon72]

I tried comet alignment with different rejection limit but I still got lots of star trails. Trails are very soft which make back ground look bright/noisy.

I am wondering how to eliminate them.
- There is rejection high image which can be used as mask. Is there a process to replace unmasked area with background?
- Generate difference between stacked images with two different rejection limit, and use this difference to remove star trails.

Has anyone had success to remove star trail residuals? What is the best way to do it?

[MikeOates]

What is probably happening here is that you have over lapping stars. You need to leave a gap in time between the exposures to stop this happening. If you have enough subs, try the alignment with alternate subs.

Mike

[blackdragon72]

It does not work either, there are still a lots of star trail residuals. I used to try min/max rejection too, thinking that it should work if star does not overlap, somehow it did not work as expected.

I am more thinking a general solution. If we have rejection map and use it as a mask, in theory we could have a script/process to replace masked area with unmasked, surrounding background value.

[oldwexi]

Hi Blackdragon72!
Had a similar challenge with comet Lovejoy. http://www.werbeagentur.org/oldwexi/gal_lti_all.html#Lovejoy

After process CometAlignment i used from the new registered
images the 1st, 6th, 11th, 16th, etc and stacked them with windsorized clipping.
Than i used the 2nd, 7th, 12th etc image and stacked these to pile 2
etc. until 5th, 10th 15th etc. i hade 5 separate stacked images where the star traces
were better removed.
I used stepsize of 5 because with my focal length of 675 mm
it needed 5 exposures to have the stars in the trail clearly separated!

Than i stacked the 5 stacks together this brought me a much better result than
stacking all images in one step.

If you have a shorter focal length your step size must be even more wide.

Gerald

[blackdragon72]

Gerald,

Nice pictures.

I did try your approach but it did not work, probably because my images did not have enough data.

I am just thinking aloud, since I have many comets photos to process.

If we can generate a star trail mask, then we could:
- either replace masked pixels with average surrounding backgrounds (I dont know if there is such script, but this will be ideal solution)
- or replace masked pixels with data from star align image using pixmath