not a bug PixInsight 1.8.9-2 build 1596 ImageSolver v6.2 stuck with high CPU

[jruuth]

When I run ImageSolver v6.2 with PixInsight 1.8.9-2 build 1596 it gets stuck on on the linked image. I am using defaults on ImageSolver. Eventually it eats 100% of CPU and I have to kill PixInsight. I am using Windows 11.

Link to image:

Integration_H.xisf

drive.google.com

Jarmo

 

[cloudbait]

When I run ImageSolver v6.2 with PixInsight 1.8.9-2 build 1596 it gets stuck on on the linked image. I am using defaults on ImageSolver. Eventually it eats 100% of CPU and I have to kill PixInsight. I am using Windows 11.

Link to image:

Integration_H.xisf

drive.google.com

Jarmo

The image header does not have enough information to determine the pixel scale, so ImageSolver just uses whatever was in there from the last time you used it. (There really needs to be a warning for that situation.) Set your pixel scale to 2" and the image solves just fine.

 

[jruuth]

Thanks! I can easily believe that the settings were not correct. I still think it should not get stuck with 100% CPU so that I need to restart PixInsight.

Jarmo

 

[cloudbait]

Thanks! I can easily believe that the settings were not correct. I still think it should not get stuck with 100% CPU so that I need to restart PixInsight.

Jarmo

ImageSolver has a history of not knowing when it is failing, and ending up in a very long, if not infinite effort to converge on a solution. Some cases have been fixed. Apparently not all.

 

[Juan Conejero]

This image has no valid image scale metadata. In these cases, ImageSolver applies its current image scale parameters. Note that this is by design, not a bug. This image can be solved without problems if valid metadata is provided (the approximate scale of this image is 2 arcsec/px).

I don't know what image scale value have you used, but you probably have managed to find one particular value that leads to a long execution time before detecting a registration error. While this is a corner case IMO (and invalid or missing image metadata is to be considered a serious error in all cases) the result is obviously undesirable. We'll try to improve ImageSolver's robustness in future versions, following a history of trying to do our work as best as we can. Thank you for reporting.

 

[cloudbait]

This image has no valid image scale metadata. In these cases, ImageSolver applies its current image scale parameters. Note that this is by design, not a bug. This image can be solved without problems if valid metadata is provided (the approximate scale of this image is 2 arcsec/px).

I don't know what image scale value have you used, but you probably have managed to find one particular value that leads to a long execution time before detecting a registration error. While this is a corner case IMO (and invalid or missing image metadata is to be considered a serious error in all cases) the result is obviously undesirable. We'll try to improve ImageSolver's robustness in future versions, following a history of trying to do our work as best as we can. Thank you for reporting.

I agree it's not a bug to use the previous scale values when they aren't found in an image. But a warning that this is happening would sure be nice!