Making a distortion model for 18mm DSRL lens, Error?

[OldSkyEyes]

Hi,

I'm trying to make a distortion model for my DSRL camera lens since the imagesolver never can't figure out the stars no matter how I try to point it in the right direction.
So I assume with a distortion model generated by the manual image solver it would work much better.

But generating the model always gives me an singular matrix error, see picture.
"Error: SurfaceSplineGenerateD(): Low-level API function error (0x0049): Singular matrix"
Am I doing something wrong or does the program not work right?

[Andres.Pozo]

I don't know what is happening. To diagnose the problem I would need both images (target and reference) and the icon with the control points.

You can use Dropbox, Google Drive or any other means to send them to me. Send me a private message if you don't want to make the download link public.

Andrés.

[OldSkyEyes]

Here you can download the Project
https://drive.google.com/file/d/1JcEtYXqO__b7__QpZMfTwQ-EIfgBPK_1/view?usp=sharing

[Andres.Pozo]

There are a couple of problems in your control point icon:

First, there are two control points in the same position (#19 & #20). Also both are invalid (magenta color). This is what causes the "singular matrix" error.
Second, there are several mismatched stars around Corona Borealis. This produces a bad distortion model and wrong astrometry.

Using the control points in the attached icon (ControlPoints.xpsm) ManualImageSolver works without problems.

[OldSkyEyes]

That works Thanks!

I see the mistakes, since I had trouble seeing which one was related to the other one.
and then it started to add points with dy > -1000 which probably introduced more error since I then reduced the dy manually to see the marker since it wasn't on the screen.
The color difference wasn't really obvious for me.

I tried to fix my own icon also instead of just using yours, and that helped me to learn to setup the tool and views to easier find wrong matches.