Problems with stacking a Galaxy in Pixinsight - Forums

Topic starter

Feb 22, 2023

Hey guys,

i am still a total noob regarding pixinsight.

last night I imaged the pinwheel galaxy for a few hours with my duo narrowband filter. the individual frames looked good but the stacked picture I get out of pixinsight is absolutely unusable.
I tried it 2 times, always using wbpp.

using app I get

i have no idea what setting could cause pixinsight to produce that mess.
maybe somebody can give me a hint what I need to change in pixinsight to get a usable result.

cs
andreas

andrea tasselli

Frequent contributor

Feb 22, 2023

I don't use WBPP and I encourage you to do likewise. Having said that it appears to me you are stacking up the un-registered lights.

A normal workflow would consist of the following:

1. Image Calibration
2. Defect Removal
3. Debayering (if OSC)
4. Sub-Frame Selection
5. Star Registraion
6. Normalization
7. Integration
8. (Optional) CFA Drizzle Integration at 1x

You can find an excellent tutorial here:
Light Vortex Astronomy - Tutorial (PixInsight): Pre-processing (Calibrating and Stacking) Images in PixInsight

Helpful

Lasse Skov

Feb 22, 2023

Hi Andreas

It's hard to say without any additional information - but have you seen Adam Blocks series about WBPP?
It's quite informative: https://www.youtube.com/playlist?list=PLAzMa9eIVQkBmzPneF8hCgqwkrxsk7CDn

Jim Raskett

Feb 22, 2023

I agree with Andrea concerning that your result does look like registration is failing or not happening.
I use always use WBPP in part or fully and have had very good success.
Being new to Pixinsight, it is necessary to understand what is going on in WBPP.
To fully understand WBPP, running each process manually to get a good understanding about what is going on is immensely helpful.
The Adam Block WBPP series as Iskov linked is excellent.
Is Image Registration selected under the “Lights” tab in WBPP?

Jim

Concise

Anderl

Topic starter

Feb 22, 2023

andrea tasselli:
I don't use WBPP and I encourage you to do likewise. Having said that it appears to me you are stacking up the un-registered lights.

A normal workflow would consist of the following:

1. Image Calibration
2. Defect Removal
3. Debayering (if OSC)
4. Sub-Frame Selection
5. Star Registraion
6. Normalization
7. Integration
8. (Optional) CFA Drizzle Integration at 1x

You can find an excellent tutorial here:
Light Vortex Astronomy - Tutorial (PixInsight): Pre-processing (Calibrating and Stacking) Images in PixInsight

I agree with Andrea concerning that your result does look like registration is failing or not happening.
I use always use WBPP in part or fully and have had very good success.
Being new to Pixinsight, it is necessary to understand what is going on in WBPP.
To fully understand WBPP, running each process manually to get a good understanding about what is going on is immensely helpful.
The Adam Block WBPP series as Iskov linked is excellent.
Is Image Registration selected under the “Lights” tab in WBPP?

Jim

I can only agree with you guys that it looks like registration failed. but in fact, It happened. I already processed like 20 or 30 images with the wbpp script and never had problems. I can try to run all the processes individually outside of wbpp but as I really like that script I would rather use it if possible.

Jim Raskett

Feb 22, 2023

Does the logfile generated my WBPP hold any clues?

Jim

Anderl

Topic starter

Feb 22, 2023

Does the logfile generated my WBPP hold any clues?

Jim

processed it a third time now just to take a look into the wbpp log file.
first time I did that but at first glance it looked normal, whole lot of stuff happening in the rejection map.
right after the galaxy I used wbpp for another picture and got a normal result.

Alan Brunelle

Feb 23, 2023

It might help to examine sub images within the folders for cosmetic correction, debayering, registration, etc. Certainly if you look at the registered images, you should be able to take two images in PI, open them, size them equally, and they should align perfectly. If they do not, then registration failed.

From what I see in your result, there seems to be a lot of shot noise still in the image. Or even hot pixels. The registration may have failed because the task is trying to align on noise rather than stars. If that is the case, your cosmetic correction may have failed to remove the bad pixels. Also, there should be a setting in the registration/star alignment that helps ignore single pixel noise and only look at bona fide stars for star alignment and then registration.

Bottom line, you have all the data you need from the process in the folder you told PI to save the data in. Its the same as if you used WBPP or did it manually. So work the problem from start to finish and find the failure point. It helps to understand how the process works, but in the end, with this issue, once solved, you will better understand what the process is (Enough comas for you?!). Otherwise, you can post some images from from the calibration folder, the cosmetized folder, the debayered folder and the registered folder. I will bet you a whole lot that your registered images do not align. If you do try to manually align them, you might actually see the features that star alignment tried to use to star align. This is because when you manually align the images one over the other, the one you are dragging with your mouse becomes partially transparent. It allows you to see instantly if your alignment is correct or if it is off. And that could give you the clue you need.

Helpful Supportive

Pranay Kanwar

Feb 23, 2023

Attaching/Looking into the wbpp log will give more insights. Usually it is in the logs/ directory of the wbpp output directory.

Concise

Anderl

Topic starter

Feb 23, 2023

Alan Brunelle:
It might help to examine sub images within the folders for cosmetic correction, debayering, registration, etc. Certainly if you look at the registered images, you should be able to take two images in PI, open them, size them equally, and they should align perfectly. If they do not, then registration failed.

From what I see in your result, there seems to be a lot of shot noise still in the image. Or even hot pixels. The registration may have failed because the task is trying to align on noise rather than stars. If that is the case, your cosmetic correction may have failed to remove the bad pixels. Also, there should be a setting in the registration/star alignment that helps ignore single pixel noise and only look at bona fide stars for star alignment and then registration.

Bottom line, you have all the data you need from the process in the folder you told PI to save the data in. Its the same as if you used WBPP or did it manually. So work the problem from start to finish and find the failure point. It helps to understand how the process works, but in the end, with this issue, once solved, you will better understand what the process is (Enough comas for you?!). Otherwise, you can post some images from from the calibration folder, the cosmetized folder, the debayered folder and the registered folder. I will bet you a whole lot that your registered images do not align. If you do try to manually align them, you might actually see the features that star alignment tried to use to star align. This is because when you manually align the images one over the other, the one you are dragging with your mouse becomes partially transparent. It allows you to see instantly if your alignment is correct or if it is off. And that could give you the clue you need.

Guess it really registered on noise. That would explain the rejection map and why the log file looked normal (at least to me)

Alan Brunelle

Feb 23, 2023

Alan Brunelle:
It might help to examine sub images within the folders for cosmetic correction, debayering, registration, etc. Certainly if you look at the registered images, you should be able to take two images in PI, open them, size them equally, and they should align perfectly. If they do not, then registration failed.

From what I see in your result, there seems to be a lot of shot noise still in the image. Or even hot pixels. The registration may have failed because the task is trying to align on noise rather than stars. If that is the case, your cosmetic correction may have failed to remove the bad pixels. Also, there should be a setting in the registration/star alignment that helps ignore single pixel noise and only look at bona fide stars for star alignment and then registration.

Bottom line, you have all the data you need from the process in the folder you told PI to save the data in. Its the same as if you used WBPP or did it manually. So work the problem from start to finish and find the failure point. It helps to understand how the process works, but in the end, with this issue, once solved, you will better understand what the process is (Enough comas for you?!). Otherwise, you can post some images from from the calibration folder, the cosmetized folder, the debayered folder and the registered folder. I will bet you a whole lot that your registered images do not align. If you do try to manually align them, you might actually see the features that star alignment tried to use to star align. This is because when you manually align the images one over the other, the one you are dragging with your mouse becomes partially transparent. It allows you to see instantly if your alignment is correct or if it is off. And that could give you the clue you need.

Guess it really registered on noise. That would explain the rejection map and why the log file looked normal (at least to me)

Log file will tell what it did in the schedule of operations. But the information can be limited and unclear. Sure it will tell you if an operation failed, but not if it did the wrong thing with the data.

With the limited information limited information you have given us, I am guessing that you failed to generate a Cosmetic Correction template and/or failed to point to it in the WBPP checkbox. Note checking the box without a template will cause an execution error, but WBPP will still complete the job. Do you see any ..._cc files in the data folders!