Plate solving keeps failing on wide angle image

Software Siril
1

I am trying to use the Astrometry Image Plate Solver under the Tools menu to solve an image but no matter how many configurations I try it always fails. Here is a link to the solved image on Astrometry.net

My imaging set up:
I am using a Canon EOS R (specs found here), and a Canon EFS 24mm f/2.8 lens stopped down to f/3.2.

https://nova.astrometry.net/user_images/15391260#annotated

There are also a couple weird nuances around this.

1st nuance - Pixel size:

The Image Information tool in Siril lists the Pixel size (microns) as 8.56, yet the specs for this camera list that as 5.34 or 5.36.

I have Binning as 1x1 and “Real pixel size” checkbox checked.

Here is what is recorded in the FITS Header for Pixel size:
XPIXSZ = 8.5600004196167 / [um] Pixel X axis size

The 2nd nuance - image dimension:

The plate solved image in Astrometry.*net list the dimension as 51.8 x 34.5 deg …

but when I put in all my specs into both Stellarium’s Sensor tool and Astronomy Tools FOV finder for imaging they both give 73.56° x 52.98°.

Finally, I tried to install the local astrometry.*net plate solver tool on my computer but both of my FOV’s were too wide for the list of options in the Index Manager for .

Is this image just too wide field for Siril to solve?

Thanks for any help!

2

Hello,

Siril internal solver probably won’t be able to solve for this FOV.
For local astrometry.net, we recommend in the documentation to use newer indexes: Platesolving — Siril 1.4.3 documentation
Looking at the log for the solve on nova, index 4116 is the one that sucessfully solved your image:
https://nova.astrometry.net/status/15132168

For the pixel size, siril relies on external libraries to read the metadata. Could you share a raw file to check what’s reported?

As for the reported FOV, i guess it has to do with the difference between the length of the segment and the arc, which, at this focal length, differ quite a lot.

Cheers

C.

3

Thanks for your response. See attached CR3 file for review.

793A9671.CR3 (11.2 MB)

4

I am not sure I understand your comment, “As for the reported FOV, i guess it has to do with the difference between the length of the segment and the arc, which, at this focal length, differ quite a lot.”

What’s the segment and arc referring to?

Thanks for your help!

5

As a follow up, I installed the local directories and the new index files via the GitHub site and it plate solved my image no problem.

1 Like
6

Checked the exif of your image. Seems the fact you’re using an EF-S lens (an APS-C lens basically) crops the FOV down to 4176x2784. But as your model is tagged as Full Frame, we compute the pixel pitch based on 36mm width, hence why the pixel size is off.
This also explains why you don’t get the FOV you’d expect, as the full sensor size is not used. With the correct pitch, I get 4176pix x 5.34um= 22.3mm
For more explanations, see this post:Solved: EOS R EF lenses reduce image resolution and aspect... - Canon Community

1 Like
7

The lens I am using is for a DSLR with a crop sensor but I am using an adapter for my full frame mirror-less camera so the actual FOV is true to a 24mm lens.

8

Never mind. I read the info at the provided link and now understand than even though an adapter will allow a lens for a crop sensor DLSR camera to work on my EOS R, it still causes a significant crop factor and loss of megapixel resolution.