›

CloneStamp Mask and SXT

Eric Gagne
in PixInsight
Hello Adam,

I am working on data from the Spaghetti Nebula that I captured with Ha/Oiii and Sii/Oiii filters.  I am learning what you teach in your Narrowband Fastrack course.  I am however facing difficulty with stars extraction.   StarXTerminator is picking up some nebulosity, same as you experienced in your Sii channel but in my case it's coming mostly from Oiii and some from Ha.

I am attaching a screenshot that shows it from the Oiii.  As you will see, they are very thin "streams" of nebulosity that are very tedious to pick up with CloneStamp and in some cases I have to ignore them because they are mixed with stars which I obviously don't want to be masked.

Is there any other technique or method to prevent SXT from picking up those parts of the nebula ?

Comments

  • Eric Gagne
    Attachment failed to upload.  Hopefully this link will work.
    Capture d’écran, le 2025-04-13 à 14.39.28.png
  • Mike Cranfield
    One technique you could try is to use clone stamp on the stars only image to remove the thin streams of nebulosity from that image. You can then form a new starless image by subtracting your clone stamp edited stars image from the original image. I say "subtract" because it looks like your images are still linear - if they are non-linear you would, of course, "unscreen".
  • Eric Gagne
    Thanks for the suggestion Mike. I wouldn’t have thought of that. I’m gonna try it and post the results.
  • BlockHeadBlockHead
    Eric,

    Mike's suggestion is excellent.
    I might suggest some creativity. What you seeing here is a result of the fact you are undersampled. You might try to temporarily upsample the image (perhaps with bicubic), do the extraction, and then downsample back. 
    Remember SXT is looking at the stars...the nebula is coming along for the ride. However the nebula really is an extrended object unlike a star PSF which SXT understands. You may get a different (better) result.
    Give it a try?

    -the Blockhead
  • Eric Gagne
    Both methods work though leading to different results.  The most promising, at least in my tests, was Mike's.  It seems to be the only one that could "save" the nebulosity.   It lead to terrible results but that is totally on me, I will have to hone my clostamping skills before I can use it.

    Adam's method works too but not as well.  I had to upsample it by 400% in order to keep nebulosity, anything lower would not make a difference but then I ended up with an unacceptable amount of noise when downsampling it back to its original size.


Sign In or Register to comment.