Stacking in PixInsight goes wrong

[jjosefsen]

Hi,

Hoping for some help here..

Im trying to learn some basic processing in PixInsight, including the PreProcessing (calibration, registrering, stacking, etc.)

Im using the only data I have, from my very first lights of M13, calibrated with a SUPER BIAS, no darks and flats yet.

When doing it in DSS I get 7 images stacked and the stacked file looks OK.

 

Switching to PixInsight it goes horribly wrong after stacking, and the stacked picture looks like the attached.

I have done the whole process in two different ways, but get the same basic result.

1. Doing all the steps manually: Calibration, Debayering, SubFrameSelection, StarAlignment, Stacking.

2. Doing everyting with the BatchPreprocessing script

I yield the same result, its like the subs aren't properly aligned, even though the process has eliminated A LOT of bad frames (from 26 to 5).

 

Link to zip with the 5 registered light: https://www.dropbox.com/s/sj2ytaa00u5qlap/GREAT CLUSTER_LIGHT_45s_400iso.zip?dl=0

 

Hope someone can give me some pointers, none of the tutorials I have been following explains why this can happen.

 

Thanks in advance stargazers!

 

[wimvb]

Most likely your image registration went wrong. If you have hot pixels, PI imageregistration can take these for stars and register the images on these.

In star detection, increase the values for hot pixel rejection and noise level / noise scale

I don't have PI open atm, but try these values

The defaults 5 0 1 0  become 5 1 2 1 (hope this makes sense)

Also, since you have star trails, change the star shape settings (can't exactly recall which setting, but read the mouse over explanations in star detection)

If everything goes ok, the background in your stacked image will show walking noise (streaks). You should include a cosmetic correction step in image calibration, to remove remaining hot pixels.

With these images, set the pixel rejections parameters in Image integration to lower values than the default 4 (low) and 2 sigma (high)

Hope this helps for now. If you need more explanations, let me know and I'll try to guide you in pixinsight.

[JorgeR]

Hello Jjosefsen, 

Were you able to resolve this? I am new to PI and getting the same results as you after running batchpreprocessing, although no errors are reported in the log. Did the answer from wimvb helped? Thanks for your reply/.

 

regards.

jorge

[steppenwolf]

If you have a lot of hot pixels that are skewing your alignment/stacking results, you could consider running the CosmeticCorrection process on your calibrated images before you star align as this will remove the hot pixels from your images taking them out of the equation.

[jjosefsen]

Honestly I just can't remember what the issue was, but I did resolve it.

I think maybe I eliminated some really bad subs with the blink process first, before registration and stacking.