Strange artifact on...
 
Share:
Notifications
Clear all

Strange artifact on mosaic


(@thedwo)
White Dwarf Customer
Joined: 3 years ago
Posts: 15
Topic starter  

Hi

I have three sets of data taken on successive nights. the first two nights they are pretty much in the same location the third being a little to one side but with more than 50% overlap.

I tried to process all in one and resulted in the attached image. You will see a striped pattern on the right hand side.

I then processed nights 1 and 2 together and the image was OK. I then processed night 3 data, which again gave a good image.

Assuming that there might have been a glitch the first time I tried the mosaic of the 3 nights, I repeated the exercise. But the image again had the stripes.

All the above processing was done on APP default settings.

Can you guide me as to what settings I perhaps should use?

Frank

GamCyg mosaic

 


ReplyQuote
Topic Tags
(@minusman)
Red Giant Customer
Joined: 5 years ago
Posts: 199
 

The stripes look like Walking Pattern Noise. Can be removed with dithering and outlier rejection.
With kind regards


ReplyQuote
(@vincent-mod)
Quasar Admin
Joined: 5 years ago
Posts: 4714
 

You did use mosaic settings or just regular integration? I'm seeing some variations and not a great blending together of the panels. Did you use MBB and LNC?


ReplyQuote
(@thedwo)
White Dwarf Customer
Joined: 3 years ago
Posts: 15
Topic starter  

Hi 

I've just repeated the processing and used dither, LNC and MMB - although I have to admit I an uncertain of the tecnicalities of these settings. The new image is virtually identical to the one I attached when opening this thread.

Rather than send the image, I think the title of the saved image may help you in seeing what actions were performed.

St-avg-8760.0s-AR_LN_6.0_3.0_none-x_1.0_LZ3-NS-full-qua-mult-sc_BWMV_nor-AAD-RE-MBB5_1stLNC_it2

I hope this is helpful

Thanks

Frank


ReplyQuote
(@minusman)
Red Giant Customer
Joined: 5 years ago
Posts: 199
 

It might be better if you can upload some raw images. From each panel 2-3 lights, and possibly Master Dark, Master Flat.
Then you can better judge what is going on with the picture.


ReplyQuote
(@vincent-mod)
Quasar Admin
Joined: 5 years ago
Posts: 4714
 

Yes that would be nice. @thedwo could you upload some raw lights (like 10 for each panel), your master bias, master dark and master flat (if you have those) to our server? Login/password is: appuser

Please create a directory with "thedwo-integrationArtefact" and upload in there. I will have a closer look then, thank you!


ReplyQuote
(@thedwo)
White Dwarf Customer
Joined: 3 years ago
Posts: 15
Topic starter  

Hi Vincent

OK but please advise how I access your server?

Frank


ReplyQuote
(@vincent-mod)
Quasar Admin
Joined: 5 years ago
Posts: 4714
 

You can click on the link in my post, the words "to our server". Then the server is loaded and you can enter the username and password.


ReplyQuote
(@thedwo)
White Dwarf Customer
Joined: 3 years ago
Posts: 15
Topic starter  

Hi Vincent

Thanks. All uploaded. 10 of each from 3 sets of lights and a Master dark. There is only a dark as you may recall from a separate thread I had problems with images taken using a different capture software.

 

I hope this enables a solution to this problem

 

Thanks again


ReplyQuote
(@minusman)
Red Giant Customer
Joined: 5 years ago
Posts: 199
 

Hey, Frank, found the problem. The Bayer CFA Pattern must be changed to GBRG in 0.
The correct RGGB is in the Fits header but apparently your images are mirrored in the Y-axis in the capture software. I came to this through the Masterdark. Because I'm using the same camera and the amp-glow is very pronounced in my upper right corner.

Here is the finished stack.

Mosaik RGB session 1 session 2 session 3 1stLNC it1 St

These black stripes should also disappear through dithering.

Bildschirmfoto von 2020 06 02 23 21 54

Quick Processing 

Mosaik RGB session 1 session 2 session 3 1stLNC it1 lpc cbg csc St

 

 

This post was modified 2 years ago 2 times by minusman

ReplyQuote
(@vincent-mod)
Quasar Admin
Joined: 5 years ago
Posts: 4714
 

Hi @minusman did you do this on the uploaded data? If so, thank you very much. This seems like that removes that pattern as well.


ReplyQuote
(@minusman)
Red Giant Customer
Joined: 5 years ago
Posts: 199
 

Yes, these are the uploaded data, but unfortunately not all lights have the appropriate masterdark and flats.

With kind regards


ReplyQuote
(@vincent-mod)
Quasar Admin
Joined: 5 years ago
Posts: 4714
 

Right, well that wasn't the biggest issue anyway it seems, is the mosaic algorithm the only thing you changed for that pattern to go away?


ReplyQuote
(@minusman)
Red Giant Customer
Joined: 5 years ago
Posts: 199
 

Yes changed the Bayer pattern.


ReplyQuote
(@thedwo)
White Dwarf Customer
Joined: 3 years ago
Posts: 15
Topic starter  

Hi Vincent/Minusman

Thanks for that info. One thing though, you say use GBRG in 0. What do you mean by 'in 0'?

I find the whole thing rather weird. Up to several months ago I was using ASICap for capture, creating darks, etc. Then these FITs were processed in APP, with no problem, including mosaics.

Then around April I started using the composite package CCDCiel (a bit like SG Pro & NINA) and that's when I started encountering processing problems, culminating in an earlier post where I found that the darks taken with ASICap and processed with CCDCiel captured lights, in APP gave a weird effect of the amp glow which showed on two places right hand side - upper and lower. This indicated that the Dark had removed the glow from the light in a different place from where is was on the light - one had been inverted. 

Previously amp glow had always been seen on upper right, indicating that when captured in CCDCiel the lights had been inverted.

I would appreciate your conclusions on this as to how I could make the settings in future.

I have to say, I am tempted to revert back to using the ASI native capture software, but CCDCiel has been useful in easily handling plate-solving and auto-focus, in one package.

A fellow astro club member, with identical setup using the ASI294MC-Pro camera, is using NINA with no problems.

Any advice appreciated.

Frank


ReplyQuote
(@vincent-mod)
Quasar Admin
Joined: 5 years ago
Posts: 4714
 

The change in software might very well have influenced your data in a sense. What some packages do is not providing more standard fits keywords or simply don't provide any so APP can't know certain things about your camera, sometimes set a flip tag on (which might be the case for you) and sometimes they change the resolution compared to the other package. Therefore it's always best to settle on 1 package to do the capturing and when you switch, making very sure the data is still presented in the same way.

0 btw is tab-0, so on the tabs in APP on the left side, with all the steps to take, 0 is the first one and it has the mosaic algorithm setting in it.


ReplyQuote
Share: