ASI1600mm PixInsight internal reflection removal

[kosborn]

Hi all,


RickS recently posted the following in a thread in the Deep Space imaging forum:





Rather than hijack that thread I thought I'd start a new one here (hopefully Rick will wander past and comment). I found the following tutorial (http://trappedphotons.com/blog/?p=616) which seems to be a good method but when the reflections are as bad as the ones I have in the attached image, it becomes too difficult. I'd be keen to hear anyone's ideas about a good solution (other than buying Astrodon narrow band filters :)).


Cheers, Kevin

[Atmos]

That is pretty useful especially for filter reflections caused by non AR issues. Sadly the diffraction pattern from the ASI1600 is different but the technique would work well for some of the cheaper filters.

[RickS]

Hi guys,

I have just been talking about this with Ben in PMs after his Horsehead post: http://www.iceinspace.com.au/forum/showthread.php?t=171958

David Ault's solution sounds like a solid one for a "simple" halo but too much work for a complex diffraction pattern like this (IMO.)

If somebody is willing to share a raw integration that exhibits the problem then I'd be happy to spend a bit of time on it to see what I can do...

Cheers,
Rick.

[ChrisV]

So is this problem specific to the ZWO camera and not the QHY163 which has AR coating? Just wondering as I'm thinking about a mono cam. Sorry to hijack Kevn.

[glend]

I believe I have a fair number of sub examples in my library but I will need to track them down. Once I discovered the problem I resorted to framing Alnitak out to avoid it. I use Baader narrowband filters, but I believe it also happens with Astrodons.

[RickS]

The reflections are off the sensor cover slip, not the camera window so the QHY version shows the same effect... or so I have read on CN.



Ben said he would get me a copy of his Oiii master which is a textbook example of the artefact, so no problem if you can't find then Glen.

[Atmos]

https://www.astrobin.com/full/321338/0/

This would be my strongest example but it’s in RGB.

[RickS]

Wow, that's very funky Colin! I'll try the mono Oiii first and perhaps work my way up to the square, multicolour RGB halos :lol:

[Lognic04]

Rick I will upload a recent stack with terrible halos.
Btw, are the halos from zwo filters or the camera itself?
Bear with the noise, it's WIP :p

[RickS]

Thanks, Logan. The halos are mostly due to the sensor (microlenses and reflections from the sensor cover glass) but I wouldn't rule out filters having some effect as well.

[Lognic04]

Here you go!
https://drive.google.com/file/d/1tOQ1aat2GbkBTOnRr8sin3bTGL_zWHcx/view?usp=sharing

[kosborn]

Happy to offer mine if you'd like. Would you like the subs or the integrated xifs?


K

[Benjamin]

Just to point out that the patterns I had around Alnitak (see below) are close to the last image on this thread:

https://www.cloudynights.com/topic/607584-halos-zwo-vs-astrodon-filters/page-2

Both are shot using both the ASI1600MM and an Astrodon OIII 3nm filter. My image is more stretched I think. An Astrodon filter (sadly) seems not to be a solution.

In processing, once a liner fit was applied to the O3, using Ha as a reference, some of the issue was quietened down somewhat.

[kosborn]

Interesting. I have ZWO filters but the narrowband ones give really bad reflections where the LRGB ones don't.


K

[RickS]

Thanks, Logan. Downloading now.



The more the merrier, Kevin :) Integrated XISF files are probably easiest.

[Atmos]

That’s pretty much it. The reflections are caused by the interaction between the microlens’ and the cover glass which isn’t AR.

This means that filters without good AR coatings are likely to be more affected (thinking ZWO narrowband) so you may get worse reflections on an ASI1600 than ASI183 on the filters themselves and not just the microlens diffraction.

[kosborn]

Here it is...https://drive.google.com/file/d/1Z80zpQOU-CCSFG1Z97IqWgVTRRoQcPLq/view?usp=sharing

[RickS]

So, Kevin and Logan have "bullseye" halos which look to me like multiple reflections from a filter external to the camera. What filters are you guys using?

Ben has a strong case of the more complex reflected microlens diffraction pattern. I had a quick play and managed to ameliorate this to a reasonable extent but not enough that I was happy.

I probably won't get a chance to do much more until the weekend and will report back on any successes... or failures :)

[kosborn]

I have a full set of ZWO filters. This particular image was the ZWO 7nm Ha filter.

[Benjamin]

Sorry to jump the gun but was having too much fun trying things! As a first step I followed Ricks advice and created a range mask, isolating the reflections and leaving the star, and then replaced the reflections with an estimate of the background (MMT). I had to soften the mask a little and still ended up with some residual halos. I did this on the three Ha, S2 and O3 tone maps and also on the Luminance (which in this version is just Ha). A start perhaps? Before and after pics below.

[AstroApprentice]

Just wondering if you have played with filter orientation? The general rule is that the reflective side should face the sky, but apparently under certain conditions the opposite may help reduce internal reflections:
https://www.baader-planetarium.com/en/blog/unmounted-filters-which-side-should-face-the-telescope/

[Atmos]

The particular issue here is mostly happening within the last 0.5mm on the sensor itself.

[Lognic04]

I too use the 7nm zwo ha.
Filter orientation hasn't helped me.

[RickS]

No problem, Ben. That definitely looks less ugly.





Are they unmounted filters? I remember reading about someone who had to mask the edges of the unmounted ZWO filters but I think that was a different issue - maybe bad flats?

[Lognic04]

Yes unmounted, but the filter wheels now come with masks so that's not the problem :(

[kosborn]

Interesting, thanks for the link. I have a field flattener that sits about 20mm in front of the filters. I was very careful to orientate the filters correctly but I might flip the narrow band ones around and see what happens.

[kosborn]

Yes mine are unmounted ZWO filters.

[AstroApprentice]

If the reflections are not exclusively due to the CMOS surface/coverslip, then Thomas' suggestion about adding spacers between the sensor and filters might also help...
https://www.baader-planetarium.com/en/blog/unmounted-filters-which-side-should-face-the-telescope/

[Lognic04]

Any luck Rick? :)

[RickS]

I had the work Christmas party tonight, Logan. I walked home and I'm not going to try driving PI tonight either :lol:

[Lognic04]

https://cdn.discordapp.com/attachments/342068044249563146/520852503458480139/unknown.png

Had a go at removing the rings with clonestamp!

[Atmos]

I’ve recently been watching Adam Blocks tutorials on masks and I’ve had an idea on how to remove circular patterns like this.

Create a circular mask that is the same size of the circular patterns. Create a star mask and subtract any stars from within that circular mask. This should create a mask that covers the rings and not the stars. Invert the mask and use a high MMT and it’ll average everything in there.

It’ll keep the Starr’s but remove the rings and diffraction spikes.

[RickS]

Had a busy weekend and it's clear so far tonight. Will get back to messing with data next cloudy night :)

The technique that Colin mentions for subtracting the bright stuff from a mask is a good one. Something like PixelMath "mask-luminance*0.75" then tweak the 0.75 value for best results... Also use HT to stretch afterwards if the mask becomes too weak.

Cheers,
Rick.

[Lognic04]

Rick what you'e describing is different to colin's method?
Colin, your method kind of worked but it just blurred it all out really :(

[RickS]

Hi Logan,

I was just talking about the part where you remove bright stars, etc. from a mask. It's a useful sub-technique.

Cheers,
Rick.

[Lognic04]

Oh I see, thanks I'll have a go tomorrow

[RickS]

My attempts at a relatively simple and quick solution with a single mask for Kevin & Logan's multiple reflection artefacts did produce an improvement but not enough for me to be happy with the result. Looks like it would need a mask for each ring to do a good job with this strategy and that's going to be time consuming...