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

Question for the experts: To do with Star Alignment on Pix Insight

astrophotography SCT
  • Please log in to reply
9 replies to this topic

#1 Ross Taylor

Ross Taylor

    Lift Off

  • -----
  • topic starter
  • Posts: 7
  • Joined: 19 Apr 2018
  • Loc: Geraldine

Posted 05 June 2020 - 12:10 AM

Hi Everyone

 

I have an interesting question for those who know something about star alignment when you're stacking imaging. I've got an 11 inch Edge and a QHY 10 CCD camera. As an example, I stacked 100 x 2 minute subs. The Star align function worked fine in Pix Insight, but having successfully run, the images are not aligned. When I stack them manually, something curious happens - if I align them all on a star in the centre of the image, then the overlapping stars going out from the centre form arcs..... so that the further from the centre, the bigger the arc that the stacked stars make. If I align the images manually using a star at a corner of the image as an alignment point, the stars heading away from that corner form arcs, radiating larger the further from the star they go.

 

It's almost as if the stacked images are slightly rotated, with rotation increasing in line with the number of subs. I'm to work out why. Any ideas would be gratefully received.....

 

Cheers

Ross

 

 



#2 albusmw

albusmw

    Lift Off

  • -----
  • Posts: 17
  • Joined: 23 May 2018

Posted 05 June 2020 - 12:19 AM

How good is your polar alignment? Think of an Alt-Az mount where rotation occurs massively - if your polar alignment from your Ra-Dec mount is too bad you will also get a slight rotation, and in a small FOV you will see it. Got it?

#3 Ross Taylor

Ross Taylor

    Lift Off

  • -----
  • topic starter
  • Posts: 7
  • Joined: 19 Apr 2018
  • Loc: Geraldine

Posted 05 June 2020 - 12:37 AM

How good is your polar alignment? Think of an Alt-Az mount where rotation occurs massively - if your polar alignment from your Ra-Dec mount is too bad you will also get a slight rotation, and in a small FOV you will see it. Got it?

Thanks - if it was polar alignment, wouldn't Star Align compensate for this?... i.e.... 120 images - all auto guided..... each has pin point stars... so there are no arcs on the individual subs.... but there are on the stacked image. 



#4 Jared

Jared

    Fly Me to the Moon

  • *****
  • Posts: 6,352
  • Joined: 11 Oct 2005
  • Loc: Piedmont, California, U.S.

Posted 05 June 2020 - 12:41 AM

Rotation between subs is usually caused by imperfections in polar alignment. PixInsight can generally account for this and should rotate all frames as required to match your reference frame.

Usually, when PixInsight fails to align the issue is hot pixels in your subs—PixInsight is trying to align those rather than aligning actual stars. Do you have any hot pixels? You’ll need to zoom in to 1:1 to see them and autostretch the image.

If you do have hot pixels left even after dark subtraction (not uncommon with short exposures and CMOS cameras) try running cosmetic correction on your subs with a relatively low sigma, perhaps 2.2. You can use real-time preview to see how far you have to go to get rid of them. If you have a subject with thirty or forty real stars and a few hundred hot pixels star alignment can fail this way and still report success since it thinks it worked.
  • Ross Taylor likes this

#5 Ross Taylor

Ross Taylor

    Lift Off

  • -----
  • topic starter
  • Posts: 7
  • Joined: 19 Apr 2018
  • Loc: Geraldine

Posted 05 June 2020 - 01:03 AM

Rotation between subs is usually caused by imperfections in polar alignment. PixInsight can generally account for this and should rotate all frames as required to match your reference frame.

Usually, when PixInsight fails to align the issue is hot pixels in your subs—PixInsight is trying to align those rather than aligning actual stars. Do you have any hot pixels? You’ll need to zoom in to 1:1 to see them and autostretch the image.

If you do have hot pixels left even after dark subtraction (not uncommon with short exposures and CMOS cameras) try running cosmetic correction on your subs with a relatively low sigma, perhaps 2.2. You can use real-time preview to see how far you have to go to get rid of them. If you have a subject with thirty or forty real stars and a few hundred hot pixels star alignment can fail this way and still report success since it thinks it worked.

That makes perfect sense. Thank you for the great answer and the idea has been echoed by someone locally with a remarkable knowledge of Pix Insight also.

 

I was actually trying to align some un-calibrated images, so that could explain it. The QHY10 camera is fairly new so one would hope that it wouldn't have too many hot pixels but you never know. Previously, I tried using batch pre-processing and used some flats, darks and bias frame but then realised that they were no good because I had changed the offset and gain settings since taking the calibration frames, so decided to use my time experimenting without calibration frames last night, then this happened, and it completely perplexed me. :-)

 

Thanks very much for your help.

Ross



#6 Jared

Jared

    Fly Me to the Moon

  • *****
  • Posts: 6,352
  • Joined: 11 Oct 2005
  • Loc: Piedmont, California, U.S.

Posted 05 June 2020 - 02:59 AM

No problem. Make new calibration files, apply them to your light frames and the alignment will probably go fine.

#7 pyrasanth

pyrasanth

    Mercury-Atlas

  • *****
  • Posts: 2,880
  • Joined: 08 Jan 2016

Posted 05 June 2020 - 03:28 AM

Also use cosmetic correction. Make sure you check the auto box initially then drill down into the auto setting & make sure the hot & cold pixel rejection options are checked. There are more advanced settings but initially auto will work fine for you.



#8 gatsbyiv

gatsbyiv

    Viking 1

  • *****
  • Posts: 815
  • Joined: 29 Mar 2009
  • Loc: Doylestown, PA

Posted 05 June 2020 - 07:09 AM

The Star align function worked fine in Pix Insight

Are you sure?  What you describe sounds like star alignment did not work (didn't change your images at all), and you just have field rotation.  If StarAlignment worked, you wouldn't be seeing this.

 

Stack your images without running StarAlignment... do you see the same effect?



#9 Jared

Jared

    Fly Me to the Moon

  • *****
  • Posts: 6,352
  • Joined: 11 Oct 2005
  • Loc: Piedmont, California, U.S.

Posted 05 June 2020 - 08:25 AM

Are you sure? What you describe sounds like star alignment did not work (didn't change your images at all), and you just have field rotation. If StarAlignment worked, you wouldn't be seeing this.

Stack your images without running StarAlignment... do you see the same effect?


I suspect he meant that it claimed to have worked—no errors—not that it actually worked. If it aligned on hot pixels, for example, that would be normal. Given the lack of calibration that seems likely.

#10 spokeshave

spokeshave

    Vanguard

  • *****
  • Posts: 2,281
  • Joined: 08 Apr 2015

Posted 05 June 2020 - 08:32 AM

I have found that when using long focal length, PI Star Alignment will mysteriously fail to align images even though the routine does not say there were any failures. I am still not sure why this happens, but I do know how to prevent it. The default RANSAC tolerance in the Star Matching section is too low for long focal lengths. For me when using 3910mm FL, I have to up the RANSAC tolerance to about 7 or so to get good registration.

 

Tim




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





Also tagged with one or more of these keywords: astrophotography, SCT



Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics