Author Topic: Star Removal (Fabian Neyer Method)  (Read 7865 times)

Offline ngc1535

  • PixInsight Old Hand
  • ****
  • Posts: 326
Star Removal (Fabian Neyer Method)
« on: 2017 November 28 20:00:11 »
Hi all,

Below is the requested tutorial(s) that demonstrates Fabian Neyer's technique of star removal. This did take a little effort to produce :) ... so I will be releasing it in parts over the next day or so. The first portion gives an introduction and shows how to create the primary star mask- the most critical thing of the entire process. There will likely be one more section. Kind of like the old TV days with cliffhangers! (Don't worry, it is just taking me time to produce the videos).

Below you can see the "before/after" of the image I was working with:

 

After I have released these tutorials, I will also eventually create a document that has a finer granularity of steps than I provide in the introduction.
A couple of other notes- these are basically first takes... and I *always* make mistakes and travel just a little off the path momentarily at times. However, this is my
style of presenting information in a way that I hope is similar to your own processing sessions at the computer.

Finally, I also intend to create a similar tutorial that tackles the challenge of the non-linear stars (you can see a few of them in the above). I will quickly go through the same steps I did above on the H-alpha data- and follow Fabian's suggestions for the non-linear ones. I didn't want to bog down the tutorial with the fickleness of the brightest stars and instead show the technique that takes care of the other 99%.

Part 1 of 2 (for this H-alpha example)
https://youtu.be/1rEBbZhsDIU


-Adam

P.S. I see I misspelled the word "yield" in my very first slide! *sigh*


Offline RickS

  • PTeam Member
  • PixInsight Jedi
  • *****
  • Posts: 1298
Re: Star Removal (Fabian Neyer Method)
« Reply #1 on: 2017 November 28 21:07:55 »
Thanks very much, Adam, and also Fabian!  I look forward to the second episode with bated breath  :D

Cheers,
Rick.

Offline ngc1535

  • PixInsight Old Hand
  • ****
  • Posts: 326
Re: Star Removal (Fabian Neyer Method)
« Reply #2 on: 2017 November 28 21:49:55 »
Hi Again,

Here is part 2 of the star removal for the H-alpha data of this thread:


Part 2 of 2 (for this H-alpha example)
https://www.youtube.com/watch?v=nD-JhXikmRE

I didn't want three parts...so this section is a little longer.

If again desired, I will also release a second version that shows how to take care of the bright (non-linear) stars on the luminance image. There is also a possibility of releasing my raw image(s).

If there are non-sequiturs in the demonstration (I recorded this at different times and states of mind)- please let me know and I will fill in the missing information.

I hope I represented Fabian's method accurately- there are some interesting/useful ideas that are all part of this technique. Replacing stars with local backgrounds at appropriate MMT scales is certainly a big takeaway. Likely there are improvements that skilled users can incorporate into what I showed.


-adam

Offline Salvatore Giambruno

  • Newcomer
  • Posts: 33
Re: Star Removal (Fabian Neyer Method)
« Reply #3 on: 2017 November 29 10:05:05 »
Great presentation(s) Adam.  I've watched the first installment, going to watch the second one.  I'll mention it here (as I did on YoutTube) for the benefit of other PI users following your tutorial: I think in the first video you forgot to binarize the second star mask.  Not sure if that will make a difference in this process further down the line . . . we'll see :)
--Sal

Offline kolec

  • PixInsight Addict
  • ***
  • Posts: 100
Re: Star Removal (Fabian Neyer Method)
« Reply #4 on: 2017 November 29 13:31:12 »
Great thank!!

kolec

Offline MikeOates

  • PixInsight Addict
  • ***
  • Posts: 278
Re: Star Removal (Fabian Neyer Method)
« Reply #5 on: 2017 December 01 09:57:10 »
Adam,

I am still experimenting, I need to get that primary star mask just right. But I am really impressed by how well it's going.

I have picked a difficult subject to remove the stars from - The Double Cluster. I have a deep Ha image (22hrs) and those two tight clusters are a challenge together with some bright halos, so I guess I need to see your next tutorial that copes with large halos better.

But I am getting pretty excited about this process, thank you so much for passing this technique on to us :)

In part 2 you made an error at about 36:15 in, you made a change, but did not say what it was, can you clarify. It relates to those tiny bits that were left and you had some larger stars still in the mask that should not have been there.

Thanks again,

Mike
« Last Edit: 2018 May 03 04:37:31 by MikeOates »

Offline ngc1535

  • PixInsight Old Hand
  • ****
  • Posts: 326
Re: Star Removal (Fabian Neyer Method)
« Reply #6 on: 2017 December 01 10:56:27 »
Hi Mike,

Yeah... I will need to look back. I made *quite a few* errors. However, as a first take hopefully not too bad.
I think what I did was to create the last mask by subtracting all other larger scales from the primary. Whatever is left over should include my little mistakes. So I didn't use the 3 pixel level...I used the "residual" mask. Anyway, if I am wrong there are a couple of very easy fixes- if the little blips remain- just run MLT with only 1 layer (1/2) excluding the first layer. This will take care of them instantly.

I will need to get a grade from Fabian at some point- and then yes, I would certainly enjoy going forward to show the non-linear star treatment.

-adam

Offline mmirot

  • PixInsight Padawan
  • ****
  • Posts: 881
Re: Star Removal (Fabian Neyer Method)
« Reply #7 on: 2017 December 01 12:28:07 »
Just,

Watched this. Great tutorial.

Thanks

Max Mirot
Sierra Vista, AZ

Offline Warhen

  • PTeam Member
  • PixInsight Old Hand
  • ****
  • Posts: 490
    • Billions and Billions
Re: Star Removal (Fabian Neyer Method)
« Reply #8 on: 2017 December 05 11:43:07 »
Well done Adam!
Best always, Warren

Warren A. Keller
www.ip4ap.com

Offline wimvb

  • PixInsight Enthusiast
  • **
  • Posts: 91
    • Wims place
Re: Star Removal (Fabian Neyer Method)
« Reply #9 on: 2017 December 25 12:37:44 »
Great tutorial, Adam. Thanks, also to Fabian.

Great presentation(s) Adam.  I've watched the first installment, going to watch the second one.  I'll mention it here (as I did on YoutTube) for the benefit of other PI users following your tutorial: I think in the first video you forgot to binarize the second star mask.  Not sure if that will make a difference in this process further down the line . . . we'll see :)
I also noted this. I think the masks need to be binarized at this level. If the mask isn't strong enough, then part of the original image will shine through, and the stars won't be removed completely. As an example: if the mask is only 50 % intensity, only half of a star brightness will be removed. (The altered image is mixed 50-50 with the original.)
Wim

AZ EQ6 GT + EQ3 PRO, SW 150P-DS, SW MN190, ZWO ASI174MM-Cool, Pentax K20D (retired),
SW ST80 + ZWO ASI120MM + Lin_guider

Offline buggs

  • Newcomer
  • Posts: 1
Re: Star Removal (Fabian Neyer Method)
« Reply #10 on: 2018 September 10 03:34:17 »
Adam,
Thanks for a great tutorial. I've been working through it but think I may have made a fundamental error (newbie and all that). I am trying this on an OSC image and all goes ok until I get to 'filling in the background sky values' where the results look awful. Just wondering if the process is only applicable to monochrome images?