Programetic image alignment tool

Leftover work from testing in October, 2025 at ODIN.

For spatial resolution measurement with Siemenstar using MTF method, we need a reference image to be well aligned with the measure sample.
It's not so easy to align the sample in the camera view so it's easier to manipulate the reference image instead.

There is a python package called pystackreg that has automatic alignment of image.
We can make a wrapper that can use scipp data structure or make sth like this.

Example Images

Here are examples of alignment.
The refernce image needs, all three types of manipulations:

• Rotation: The pictures are rotated from how the camera is mount or sample is mount.
• Scaling: Based on the camera setting, the picture can be zoomed in or out.
• Translation

Reference Image

Measured Image

Start by reviewing the pystackreg package and how scipp represents image data. Define the wrapper boundary and verify it can align the reference and measured images through rotation, scaling, and translation using the provided examples.