PixInsight Forum

PixInsight => Image Processing Challenges => Topic started by: chandrainsky on 2019 December 27 07:52:16

Title: Error in Star Alignment
Post by: chandrainsky on 2019 December 27 07:52:16
Trying to star align images of Eskimo Nebula shot with 8" f/10 SCT. The number of stars is limited but certainly there are more than a dozen stars in all images. But I keep getting the error messages:

*** 0 star pair matches found - need at least eight matched stars

and

*** Error: Unable to find an initial set of putative star pair matches

I tried to play with various star detection parameters but nothing seems to work.
The last set of parameters that I tried is provided in the attachment.

Could you help me figure out what could be wrong? Uploaded three of the images from the set I am trying to align in Google Drive:

https://drive.google.com/open?id=1zQilcyq2EzeRMxp9VKSi78HtB_V1_F7a
Title: Re: Error in Star Alignment
Post by: bulrichl on 2019 December 27 09:28:28
These 3 image are  aligned flawlessly applying the default settings of the StarAlignment process (PixInsight 1.8.8-3):

<No error message>,
===== StarAlignment: 3 succeeded, 0 failed, 0 skipped =====

Please just reset the StarAlignment process, select the reference image, select the target images and apply the process.

Appended is the comparison of your first image and the integration of the 3 aligned images (same STF applied) and the default settings of StarAlignment that I used.

Bernd
Title: Re: Error in Star Alignment
Post by: chandrainsky on 2019 December 27 20:48:39
Thank you for the quick response. Looks like there is something else going on here. When I had tried the alignment earlier, I had included a total of 144 images of various exposure times (15 sec to 60 sec) with different filters  (H Alpha, O3, S2). I am using ZWO1600 Mono Cooled cam.  I had changed various settings and after your response I used the default settings and only 7 images were aligned.

To make sure that I am aligning 'similar' messages, I tried aligning with the following results. Please note that I have modified Max distortion and Descriptor Type 


With only H Alpha images - 60 sec exposures - Max distortion 0.33 - Pentagon for similarity

===== StarAlignment: 18 succeeded, 12 failed, 0 skipped =====



With only H Alpha images - 60 sec exposures - Max distortion 0.33 - Triangle Similarity

===== StarAlignment: 28 succeeded, 2 failed, 0 skipped =====



With all files (H Alpha, O3 and S2) - Max distortion 0.50 - Triangle Similarity

===== StarAlignment: 19 succeeded, 132 failed, 0 skipped =====


To see why the number of aligned images is reduced when images with different filters were tried, I next tried alignment of only O3 images as follows:

With only O3 images - 15 sec exposure time - Max distortion 0.50 - Triangle Similarity

===== StarAlignment: 2 succeeded, 60 failed, 0 skipped =====

So, what's happening? Why number of aligned images is changing when images with different filters are included even though the reference image is the same?
When only H Alpha images are aligned, more images of H Alpha are aligned compared to when trying alignment with images taken through different filters. Why some of those H Alpha images are getting "non-aligned"? Am I missing something basic?
Also, I see that in many cases PI is able to find 3 to 5 star pairs but not the minimum of 6 required. Tried changing a few parameters to relax the alignment process but nothing seems to work.

Here is the link for four images taken with O3 filter which were part of the alignment process:

https://drive.google.com/open?id=1CZi59GstUSUOLdXAkxoGHFwKyt6h6zLl
Title: Re: Error in Star Alignment
Post by: dave_galera on 2019 December 28 07:46:13
Think you may have to upload a full set of images so it can be investigated properly
Title: Re: Error in Star Alignment
Post by: bulrichl on 2019 December 28 10:11:00
I was able to successfully align 6 images (of 7) using the first Ha image as reference. However, one OIII image failed: "L_OIII_2019-04-05_20-00-18_15s__-12C_NGC2392_Bin1x1.fit". This image obviously has worse focus than the rest of the images.

In the section 'Star Matching', 'Triangle similarity' was chosen as Descriptor type, and the settings in the section 'Star Detection' were as follows:

Detection scales:   5
Noise scales:      2
Hot pixel removal:   2
Noise reduction:   0
Log(sensitivity):   -3.00
Peak response:      1.00
Maximum distortion:   0.500
Upper limit:      1.000
Compute PSF fits:   Never

My guess is that all of your images are exposed way too short (only 15 s!), this is not sufficient, particularly with the OIII filter. So a longer exposure time will be needed.

Furthermore the images are not calibrated with a MasterDark, so hot pixels might disturb the registering process additionally.

Bernd