Batch Preprozessor issue: Bright stars center defect

sonixx

Member
Oct 6, 2019
6
0
All,

I always used the manual calibration and stacking process without issue. But with the new weighted batch script I wanted to try this.
However on bright stars the center is always defect like in the attached picture. It is the unstretched raw integration.
I played a bit also with the normal batch preprocessor, with same results.
As the rejection algorithm I always use the winsorized rejection algorithm with standard settings which works fine.
To make another test I did also trial runs without calibration files, also same result here.

The camera is an OSC ASI071MCPRO.

Any Ideas?

Best regards,

Stephan
 

Attachments

sonixx

Member
Oct 6, 2019
6
0
Thank you for your quick reply. Actually it was activated first. I rerun the script with deactivated drizzle with a similar result.
 

ngc1535

Well-known member
Feb 1, 2014
481
46
It would be difficult to say what is going on without examining the data in detail (and the settings you used).
The saturated stars must, at some level, be in the raw data. The combination of weights and rejection will only emphasize some fraction of data/values that already exist. For example, you would need to use the same rejection algorithm and parameters that WBPP auto selects for you (I am assuming you did not force a particular method).
-adam

P.S. I would be happy to work with you on examining the data. Find me via private messaging if you wish.
 

sonixx

Member
Oct 6, 2019
6
0
I did some other trials, without success. Thanks you for your offer, I will send you a PM.
 

ngc1535

Well-known member
Feb 1, 2014
481
46
So...I looked at your images. The stars that concern you are saturated in the original data.
That represents that start of how artifacts can happen.
Bayer matrix saturation is tricky since it can happen more so in one color rather than the others.
You did not provide the image that you said is OK "Without issue."
I guess I need to see that one, since everything should still be saturated there as well.
But the summary is... this is what will happen to your saturated stars. (but feel free to pass along the combined image that does not show the issue)
-adam
 

sonixx

Member
Oct 6, 2019
6
0
Adam,

thank you very much. I can confirm that I only get these artifacts with both batch pre processing scripts.
In the folder I shared I added now the two following files:
-A manual image integration made from the registered files that came from the weighted batch pre processing script
-The integration icon with the settings used to integrate those images

As can be seen the stars are fine. In my view the automated scripts do have an issue that produces these artifacts, at least with my camera.

So maybe additional settings are necessary for the automated scripts for correct results?

Best regards,

Stephan
 

Juan Conejero

PTeam Member
Sep 2, 2004
7,364
58
56
Valencia, Spain
pixinsight.com
Hi Stephan,

These artifacts are the result of wrong pixel rejection settings.

The image integration feature of the WBPP script is just a convenience feature, which has been included just to get a rough preview of the achievable image. WBPP's integration should never be used for production purposes. Image integration must always be performed manually to optimize SNR and pixel rejection parameters.

Here is a verbatim copy of the warning message shown by WBPP when you enable its image integration feature:

You have selected to perform an integration of light frames with this script.

Please keep in mind that the light frames integration functionality of this script is just a convenience feature, which we have included to let you take a quick look at the final image. It will give you an idea of the achievable image, but in general, it will not provide an optimal result. In most cases the integrated result of this script will be rather poor, compared with the image that can be achieved by optimizing image integration parameters.

Image integration is a critical task that requires fine-tuning. Our ImageIntegration tool allows you to find optimal pixel rejection parameters to maximize signal-to-noise ratio with the appropriate rejection of spurious image data. In general, this requires some trial-error work that can't be done automatically from this script.
The old BPP script is now obsolete and will be removed from the standard distribution in the next version of PixInsight.
 
  • Like
Reactions: sonixx

sonixx

Member
Oct 6, 2019
6
0
Hi Juan,

thanks a lot, my bad. I actually clicked away this warning without thinking about too much.

Is there anything I need to do so the weighting of the registered files make it into the final image using the image integration tool?

Thanks a lot!

Stephan
 

Juan Conejero

PTeam Member
Sep 2, 2004
7,364
58
56
Valencia, Spain
pixinsight.com
Is there anything I need to do so the weighting of the registered files make it into the final image using the image integration tool?
When the Generate subframe weights option is enabled, WBPP generates FITS keywords with the name 'WBPPWGHT' and values equal to computed weights for each image. So you just have to select the following in ImageIntegration:

Weights = FITS keyword
Weight keyword = WBPPWGHT

to use weights computed by WBPP.