WBPP Local Normalization Error "No reference image has been defined"

[lpeterson1]

Hi PixInsight,

I have been getting this error from WBPP on some recent images:

*** Error: PSFScaleEstimator::EstimateScale(): Internal error: No reference image has been defined

I have been experimenting with higher camera gain settings, so there are more saturated stars in these images. Would that be a possible cause? Is there a work around or a recommended WBPP setting to change?

Thank you!

 

[robyx]

HI @lpeterson1,

Could you share the log file? I would check the in detail execution to see what's the issue precisely.

Robyx

 

[lpeterson1]

Hello @robyx,

It is attached.

Best regards,

Lee

 

[lpeterson1]

I may have found a fix. I dug in the log file a bit and saw another error message about not being able to determine a scale for the images. I changed the scale analysis to "multi scale" and local normalization processes the images without error. However, it takes a couple hours, but it is worth the wait.

Does this make sense or is there perhaps another, simpler fix?

 

[robyx]

Hi Sorry I didn't follow up this thread.
Would it be possible to share the data? I need to reproduce the issue to verify exactly where the problem is.

The error reveals an issue occurred while LN is measuring the image's star flux, switching the Multiscale solves because it means adopting a method based on wavelets and not on star fluxes.

Let me know if this is possible, thanks!
Robyx

 

[lpeterson1]

Yes I would be happy to do that. Can I email you a dropbox link outside of this forum?

 

[lpeterson1]

Actually I figured out that the link is encrypted so here it is:

pixinsight_test_data

Shared with Dropbox

www.dropbox.com

I will delete the link as soon as you let me know you received the files so that I can clear out the space on my dropbox. It is about 7-8 GB.

 

[bennich]

I am dabbling around with Local Normalisation as well and are seeing the same error when I use WBPP. I have been running both against my full set of data, and against a subset of data.
Logfile attached here.
Let me know if it would be helpful to share the data as well.

 

[bennich]

Screenshots from Diagnostics.

 

[jbernstein1]

I am having a similar problem. When I use flats from a 'flat panel', I get the error. When I use flats from an iPad, I don't get the error. Here is the log file when it failed.

 

[afrantzis]

Hi all,
Same problem for me.
"Error: PSFScaleEstimator::EstimateScale(): Internal error: No reference image has been defined"
and then "Error: Unable to compute a valid scale estimate"

I have to note that with the dataset that I had the problem, I also had for the first time quite some trouble to registered the frames. I succeeded in doing so at the end by playing with the parameters.

My images (C14 HD with 0.7x reducer = 2,737 focal length) concerned the bright but "small" planetary nebula "The ghost of Jupiter". I also tried Multiscale with no success.

I have attached one of the LN ref file created by the process (reduced in size as .jpg) and the log file, in case someone can help.

--> Update: By playing with various parameters I succeeded to complete LN and Integration of 29 out of 41 frames. I observed that it is always ch2 (I suppose this is the blue channel) the problematic one and no valid PSF could be produced for the discarded frames.

Best wishes,
Alexandros