The image of the witch head nebula below is from my dark sky site at Chiricahua Sky Village. It is a stacked image from WBPP in Pixinsight with the base stretch from STF. The artifact at the top left of center seems to be light flaring in from Rigel. Does anyone here have experience in removing an artifact of this nature ?
Details:
AT60EDP, ASI2600MC Duo, AM5, No filters,
Aperture 60mm, FL 300mm
36 × 300sec light, 10 × 300 sec dark, 30 x bias, 30 x flat.
📷 Test1.jpeg
I can’t offer much help here, but I had a similar issue. I think the path for this was that Rigel was just off the sensor (300mm f/l scope with ASI294MM) and was focused on a shiny part of the focal plane around the chip. The reflected light then bounced off of each filter and onto the sensor. I tried to map this out by stepping away from Rigel, but could not keep the main subject in frame. The best test would be to remove a filter and see if the reflection goes away but the scope is at a remote site. On the positive side, I managed to give the witch some amazing hair!
Richard Kelley · Jan 23, 2026 at 12:27 PM
I can’t offer much help here, but I had a similar issue. I think the path for this was that Rigel was just off the sensor (300mm f/l scope with ASI294MM) and was focused on a shiny part of the focal plane around the chip. The reflected light then bounced off of each filter and onto the sensor. I tried to map this out by stepping away from Rigel, but could not keep the main subject in frame. The best test would be to remove a filter and see if the reflection goes away but the scope is at a remote site. On the positive side, I managed to give the witch some amazing hair!
Thanks. I am not using filters for this image.
andrea tasselli · Jan 23, 2026 at 10:36 AM
Not easy, if possible at all, given how deep the flare is. Blotting it out isn't really feasible given the surrounding nebulosity. You might try to move Rigel in the frame, take a number of frames with the same exposure and use the portion of background unaffected by the flare to replace the one affected by the flare , which means registering the latter with the former. Once you get there you add them side by side and clone stamp the flare out using the background of the image unaffected by the flare. Then you can crop out the extra image added to the side of the original. All of this assumes PI.
Thanks. I guess I should try and figure out how it happened.
Ian McIntyre · Jan 23, 2026 at 02:06 AM
Use GAME script, which is a freeformish-ish selection tool. https://youtu.be/mG3EvOtQ4cY
On the other hand, my unsolicited opinion is I rather like it the way it is. Rotate to portrait and the witch is under attack from an unseen spellcaster. Fits her expression.
Thanks. Never had much luck with the GAME script. I’ve been working with Blemish Blaster from SETI Astro.
Jim Raskett · Jan 23, 2026 at 01:46 AM
You could try removing the stars and clone stamping the remaining flare on the background image.
Thanks. After using StarXterminator some of the flaring makes it onto the star image as well.
Thanks to all for your thoughts.
Let me just add that I do have some experience in removing artifacts within Pixinsight using GAMES and, more recently, Blemish Blaster.
The problem here is that the artifact is about the brightest object in the image and it penetrates quite far into the field. Also there is a lot of background nebulosity in the starless image which complicates removing the artifact.
The data looks quite promising otherwise and I’m reluctant to abandon it.
PI really isn’t the best program for what you want to do. Your best shot would be to use a photo editing program as it’s designed to do this kind of work. Affinity is free and it’s great for putting the finishing touches on images that come out of PI or Siril or whatever. The time involved in learning how to tackle this problem in a photo editor will be time well spent.
I’ll take a look.
Thanks.
After struggling with this for several days, Tom Fleming (https://app.astrobin.com/u/TomArizona#gallery) was kind enough to work on removing the artifact for me. The final result, a slight tweaking of Tom’s work, is shown here: