Image classification (transformation)

So, I'm at the finals of my undergraduate and searching/writing my thesis, about Image Processing. In this point, I got some photos captured by drone, separated in 5 different color bands. At the beginning, using cpselect function, I registered the points between the fixed image and other moving images (because the are some distortions between the captures). Now I need to classificate the transformations among this images (rigid, non-rigid, etc.) using the dots I registered with cpselect. However, I could find large number of articles about image registration and classification, but none of them joins both or gives some way/rule or algorithm to classificate this images. Can you help me with this?

Summarizing, I need some function/algotithm that takes this registered points, compare between the images, and (with any rule or method) tells what kind of transformation there is between the images.

After thar, my job will be to create another code to resize this images, according this classifications.

Thank you so much!