Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Anomalies found after using PixInsight WBPP

  • Please log in to reply
18 replies to this topic

#1 Rmorgan

Rmorgan

    Ranger 4

  • -----
  • topic starter
  • Posts: 324
  • Joined: 01 Dec 2009
  • Loc: Polk City Florida

Posted 14 October 2021 - 09:18 AM

The attached file is a zoomed part of a file that was created from the WBPP in PixInsight, no other processing was done.  I'm trying to figure out the source of the anomalies pointed out by the arrows.  Would these be bad pixels and how would I get rid of them showing up?

 

Thanks!

--Rob

Attached Thumbnails

  • Dither2.jpg

Edited by Rmorgan, 14 October 2021 - 09:23 AM.


#2 terry59

terry59

    Voyager 1

  • *****
  • Posts: 10,535
  • Joined: 18 Jul 2011
  • Loc: Colorado, USA

Posted 14 October 2021 - 09:23 AM

no file but I'm going to guess that CC will help


  • Rmorgan likes this

#3 chanrobi

chanrobi

    Viking 1

  • -----
  • Posts: 960
  • Joined: 02 Sep 2019

Posted 14 October 2021 - 09:34 AM

Seems like thats way too many to be bad pixels, and if it was you'd see that same fixed pattern on multiple frames?



#4 ryanha

ryanha

    Viking 1

  • *****
  • Posts: 920
  • Joined: 05 Aug 2020

Posted 14 October 2021 - 09:37 AM

There are a lot of steps that go into WBPP.  Flat and dark calibration, then registration, and integration.  

 

When I see stuff like this, first step is to blink the raw subs and see if I find the artifacts in the raw subs, then do the same with the calibrated subs.

 

Hopefully you just have a bad calibration master (flat or dark) and you can fix that and redo.  

 

To me this looks like stacked fixed pattern noise (fixed position on the sensor).

 

Are you dithering?  Maybe need a more aggressive dither if you are?

 

--Ryan


  • 925chicago likes this

#5 925chicago

925chicago

    Explorer 1

  • *****
  • Posts: 96
  • Joined: 09 Aug 2018
  • Loc: Mokena, IL

Posted 14 October 2021 - 09:40 AM

This reminds me of the walking noise with the DSLR I used to see before I included dithering in my imaging. Do you dither?


  • Poynting likes this

#6 ryanha

ryanha

    Viking 1

  • *****
  • Posts: 920
  • Joined: 05 Aug 2020

Posted 14 October 2021 - 09:42 AM

If it turns out that it is fixed pattern noise (which you can prevent from impacting your stack in future by more aggressive dithering) you may still be able to clean up what you have with cosmetic correction.

 

Here is a thread from PI forum on CosmeticCorrection in WBPP, might be a good start.

 

https://pixinsight.c...p-please.16620/

 

--Ryan


Edited by ryanha, 14 October 2021 - 09:45 AM.

  • Rmorgan likes this

#7 Rmorgan

Rmorgan

    Ranger 4

  • -----
  • topic starter
  • Posts: 324
  • Joined: 01 Dec 2009
  • Loc: Polk City Florida

Posted 14 October 2021 - 09:54 AM

I compared a single light and it looks like bad pixels?  Not sure why they look like streaks.  Would dithering need set in PHD or SGP or both?

 

--Rob

Attached Thumbnails

  • Dither3.jpg


#8 925chicago

925chicago

    Explorer 1

  • *****
  • Posts: 96
  • Joined: 09 Aug 2018
  • Loc: Mokena, IL

Posted 14 October 2021 - 10:18 AM

You can set dithering up in SGP alone - it will take care of it for you via communication to PHD. I believe the elongation of the stars in your stack (and single light) indicates a possible RA backlash or a polar alignment error, and that's why the bad pixels look like streaks. They are roughly in the same direction as the elongation of the stars.

 

I would start by setting the aggressiveness in SGP to a medium value for a start.



#9 terry59

terry59

    Voyager 1

  • *****
  • Posts: 10,535
  • Joined: 18 Jul 2011
  • Loc: Colorado, USA

Posted 14 October 2021 - 10:24 AM

I compared a single light and it looks like bad pixels?  Not sure why they look like streaks.  Would dithering need set in PHD or SGP or both?

 

--Rob

Cold and hot pixels respectively



#10 Rmorgan

Rmorgan

    Ranger 4

  • -----
  • topic starter
  • Posts: 324
  • Joined: 01 Dec 2009
  • Loc: Polk City Florida

Posted 14 October 2021 - 10:54 AM

Looks like I get another clear night tonight so thanks for the suggestions.

 

--Rob



#11 terry59

terry59

    Voyager 1

  • *****
  • Posts: 10,535
  • Joined: 18 Jul 2011
  • Loc: Colorado, USA

Posted 14 October 2021 - 11:14 AM

Looks like I get another clear night tonight so thanks for the suggestions.

 

--Rob

Before changing anything try adding CC to your WBPP run. CC is easy to add when removing hot and cold pixels



#12 Rouzbeh

Rouzbeh

    Apollo

  • *****
  • Posts: 1,402
  • Joined: 28 Jun 2006
  • Loc: Canada

Posted 14 October 2021 - 11:15 AM

Yes I think Cosmetic correction should help a lot.



#13 Rmorgan

Rmorgan

    Ranger 4

  • -----
  • topic starter
  • Posts: 324
  • Joined: 01 Dec 2009
  • Loc: Polk City Florida

Posted 14 October 2021 - 12:07 PM

Tried adding CC to WBPP but didn't seem to help.  Went with the default settings as mentioned in the link.  It's odd because all the individual files under Calibrated, Cosmetized, and Debayered all look fine.  Must be something in them.  Also tried CC on the lights and then ran the new lights through WBPP.  Same result.

 

--Rob



#14 terry59

terry59

    Voyager 1

  • *****
  • Posts: 10,535
  • Joined: 18 Jul 2011
  • Loc: Colorado, USA

Posted 14 October 2021 - 12:34 PM

Tried adding CC to WBPP but didn't seem to help.  Went with the default settings as mentioned in the link.  It's odd because all the individual files under Calibrated, Cosmetized, and Debayered all look fine.  Must be something in them.  Also tried CC on the lights and then ran the new lights through WBPP.  Same result.

 

--Rob

Can you make a sub available. I'd like to try 



#15 SeymoreStars

SeymoreStars

    Gemini

  • *****
  • Posts: 3,408
  • Joined: 08 May 2014
  • Loc: Pennsyltucky

Posted 14 October 2021 - 12:37 PM

I compared a single light and it looks like bad pixels?  Not sure why they look like streaks.  Would dithering need set in PHD or SGP or both?

 

--Rob

Darks should take care of the hot pixels.



#16 Rmorgan

Rmorgan

    Ranger 4

  • -----
  • topic starter
  • Posts: 324
  • Joined: 01 Dec 2009
  • Loc: Polk City Florida

Posted 14 October 2021 - 01:57 PM

Here is a link to 4 files sets. Each set is zipped; DarkFlat, Dark, Light, Flat.

 

https://drive.google...TLl?usp=sharing

 

I'm uploading now, my internet is a bit slow so 5 more mins

 

 

--Rob



#17 terry59

terry59

    Voyager 1

  • *****
  • Posts: 10,535
  • Joined: 18 Jul 2011
  • Loc: Colorado, USA

Posted 14 October 2021 - 02:44 PM

Here is a link to 4 files sets. Each set is zipped; DarkFlat, Dark, Light, Flat.

 

https://drive.google...TLl?usp=sharing

 

I'm uploading now, my internet is a bit slow so 5 more mins

 

 

--Rob

The frames I looked at had hot pixels but I didn't see the cold pixels you showed....different data?  CC worked great on the hot pixels. Here is a before and after

Attached Thumbnails

  • Capture.JPG
  • Capture1.JPG


#18 Alex Ranous

Alex Ranous

    Sputnik

  • -----
  • Posts: 35
  • Joined: 05 Sep 2007
  • Loc: Redwood City, CA

Posted 14 October 2021 - 02:46 PM

I think what's happening is that the hot pixels in your darks have a greater value than the hot pixels in your lights, so when the darks are subtracted, the hot pixels become 0.  I zoomed in on one of your lights that showed the hot pixels and then opened the master dark and zoomed it to the same level.  Blinking between the frames shows the hot pixels lining up, but when I check the value of the respective pixels.  The pixel I looked at had a value of 2946 in your light and 3352 in your dark.  The fits headers show your dark was exposed at the same duration and length.  Did you use an offset when you took your images and was it the same for both the darks and the lights?  A mismatch there could explain why your darks are matching your lights.

 

 

 

The problem with the instructions for CosmeticCorrection in that link is that it was only setup to remove hot pixels.  You need to select a cold sigma under auto detect as well.  When I did that, the resulting master light no longer had the dark streaks.

 

Here's the result with the right one only with hot sigma checked, and the left with both hot and cold.

 

Screen Shot 2021-10-14 at 12.48.04 PM.png


Edited by Alex Ranous, 14 October 2021 - 02:51 PM.

  • limeyx likes this

#19 Rmorgan

Rmorgan

    Ranger 4

  • -----
  • topic starter
  • Posts: 324
  • Joined: 01 Dec 2009
  • Loc: Polk City Florida

Posted 14 October 2021 - 03:20 PM

With the camera's ascom driver I did set the Offset to 10.  I thought it would persist throughout SGP but I'll check.

 

Ran WBPP again and setting both the hot and cold seems to have correct it. 

 

So tonight I'll set a dither, check the offset, check the pa, and try another set of lights.

 

Thanks for talking the time to help!

 

--Rob


  • SeymoreStars and ryanha like this


CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics