Image Solver problems

[ppistor]

In the 1457 times, I succesfully resolved an extremly wide image (Focal length: 20 mm; covers area from Orion to Plejades). Quality was acceptable.
With the 1473 version, I expected quite am improvement by running this image through the optimisation option of ImageSolver. However, instead of
slightly improving the starting point (5 26 15 (RA), 16 46 03 (DEC)) it reslted in image center coordinates about 17h (RA) and 72° (DEC).

I made another attempt to do image solution from scratch (Dynamic Alignment, Manual Image Solver,  final optimasation by Image Solver). Here, step two failed.
In spite of having 122 control points (fairly well distributed over the image), I received the following error message:
"Error: SplineWorldInformation::InitializeSpline(): PointSurfaceSpline::Initialize(): Insufficient Data ".

Peter

[Juan Conejero]

Bug confirmed. This one requires a new version of the core application, which I am preparing now as a regular update. Sorry for the inconvenience.

Besides that, have you tried to solve the image with the ImageSolver script?

[ppistor]

Yes, I did. Same inaceptable results

[Juan Conejero]

Could you please upload the image (dropbox, etc) so I can investigate the problem?

[SteveTeel]

I notice that in the 1474 release the Epoch parameter is missing from the Image Solver script. If the image comes from a data source that does not record the DATE-OBS keyword, one must be sure to enter it using the FITS Header process.

[Juan Conejero]

Yes, this is by design. A warning message is now shown when trying to solve an image without knowing an actual acquisition date, either automatically gathered from existing metadata, or entered manually. With the increased accuracy of the new distortion modeling algorithm that we have implemented in the ImageSolver script, computing accurate star proper motions is critical.

[ppistor]

With the last PI update, the sequence DynamicAlignment + ManualImageSolver worked. When annotating the resulting image I observed an annotation
quality I rarely experienced with extremely wide images. Thanks for that.
Alas, when running the result thru ImageSolver, it refused to provide a solution. I am uploading test images as soon as possible.
Peter.

[ppistor]

Here are the images I have been asked to send in:

https://www.dropbox.com/sh/d2vc9ttd3atql9l/AAAFxyi-Od6mykDlcMId2Z7la?dl=0

Peter

[Andres.Pozo]

I can confirm that there is a problem in the implementation of the Stereographic projection. I have uploaded a fix to the repository.

Meanwhile you can try to repeat the process using other projection (i.e. gnomonic).

[ppistor]

I tried the ImageSolver script on
I4421_4740_DbeBnCcMLtHtHDR_CT_clone_registered.xisf
(see https://www.dropbox.com/sh/d2vc9ttd3atql9l/AAAFxyi-Od6mykDlcMId2Z7la?dl=0)

Using "Gnomonic", I tried 2 options:
_Using "optimisation only", I got a reasonable result, but errors in declination that were clearly visible after annotation.
_Without  "optimisation only", image solution failed.

I should note the observation date: '2019-03-23T21:00:00:00.000'.

Peter.

[Juan Conejero]

I can confirm that there is a problem in the implementation of the Stereographic projection. I have uploaded a fix to the repository.

Meanwhile you can try to repeat the process using other projection (i.e. gnomonic).

This is now fixed with a regular update, which we have just released for current 1.8.6 versions.