Astroart Forum

They are present throughout the whole frame.

Thinking that they might be Hot Pixels, I applied the hot pixel filter to each one of the 20 x 5mins subs before dark frame subtraction, aligning and stacking (Sum).

I Suspect that they are hot pixels, which have been elongated during guiding, and thus being shifted in the frame relative to each subs, but neither dark subtraction or the hot pixel filter removes them

This is obviously just a small section of the image, and highly stretched so the problem can be more easily seen, and to this end, I have also blurred the background.

I am using AA8

Dave_S wrote: ↑

11 Feb 2022, 23:05

They are present throughout the whole frame.

Thinking that they might be Hot Pixels, I applied the hot pixel filter to each one of the 20 x 5mins subs before dark frame subtraction, aligning and stacking (Sum).

I Suspect that they are hot pixels, which have been elongated during guiding, and thus being shifted in the frame relative to each subs, but neither dark subtraction or the hot pixel filter removes them

This is obviously just a small section of the image, and highly stretched so the problem can be more easily seen, and to this end, I have also blurred the background.

I am using AA8

pixels.jpg

I'd try a sigma plus sum and run the hot pixel filters after the stack.

Ok, I'll give that a try.

I have just been averaging the dark frames.

I seem to remember reading somewhere that no form of pre-processing should be applied to these CMOS dark frames. I wonder if that means doing nothing to them, i.e selecting "None" in the option menu, apart from stacking them?

Don't remember seeing this problem back in my CCD days.

Tried that, but no different.

I might try using Astro Pixel Processor, if there is a demo version available.

I cannot see any trace of them in the individual subs, albeit if they are hot pixels, they would be no more than specks in an individual sub.

It is not until they are stacked (SUM) that they become evident, either before, or after dark frame subtraction.

The image shown here is aligned, stacked (SUM), but no debayer or dark frame subtraction.

The image is heavily zoomed to make then more visible.

Ok, I have downloaded a 30 day trial version of APP, but having had a quick look, it is a whole new learning curve, and not particularly user friendly. A headache in itself. I thought PixInsIght was bad enough

I'm hoping that I can resolve this issue with AA8, and can continue using it.

A bit more info.

Aligning and stacking (SUM) the light frames, no dark frame subtraction, then applying the Hot Pixel filter using the Bayer Matrix option (100% strength) to the stacked light frame, reduces the intensity of these "hot pixel worms", but does not remove them completely.

Is there an option to open all of the light frames, and batch apply the Bayer matrix hot pixel filter to them, prior to taking them into the pre-processing window. Applying the filter one at a time to each light frame, then saving it, is a 'pain'

I know that there is an option under pre-processing, to apply a hot pixel filer, but you cannot choose which type (single, group, or Bayer Matrix).

Success

Stacking the light frames using Sigma + Sum, then applying average Darks, removes these "hot Pixel worms" from the final image.

I don't profess to understand the mechanism involved, although I'm sure Fabio will, but it works.

Whether using Sigma + Sum has any negative impact on the image, as opposed to just using "SUM" I don't know?

Hi Dave, you may also solve with the Preprocessing Filter:

Hot pixels groups, strength = 50.

See attach, Fabio.

Hi Fabio

Yes, that does indeed also work.

Is there any preference for one method over the other, or is either method as good as the other?

Dave

Hi Dave, Sigma is usually preferred because it will correct at the same time other defects, like satellite strays. It's slower by the way and it needs several images with random shifts.

Filtering for Hot Pixels can be better if the result image is to be used for astrometry and photometry, unless the image is undersampled (so, with stars of 1-2 pixels).
In this case, Strength = 50 may be too high, try 30 down to 10 to find the minimum value which just corrects the bad pixels.

About Sigma+Sum vs Sigma see also: viewtopic.php?f=2&t=616

Hi Fabio

Thanks for that, it is useful information.

The fact that Sum + Sigma will deal with satellite 'streaks' is a very useful feature indeed, as these are a growing problem, year on year.

Being slow is not a problem, as being retired, I have all the time in the world.

I print off all these useful bits of info, and put them in my AA8 'ring binder' for future reference.

Thanks again

Ciao

Dave