How can I match two similar edge images of different dimensions, and superimpose their original images correctly?

I have two pictures of the same scene in different light profiles that I would like to superimpose so I know the coordinates to transfer markers between the images. They are of different dimensions, colour and resolution so I have used an edge detector to obtain two similar edge maps of the images.

I would like to know a robust method for matching these edges on each other, or another way of superimposing the original images accurately.

I have a lot of pairs like this and the location of the zoomed image within the other is not the same each time.

[Nick Hobbs]

You may find some of the following resources useful. MATLAB is capable of matching features and can also be used for edge detection. There might be a way to use this information to allow MATLAB to perform the matching that you require.

Answer 1

[Image Analyst]

You might try Hu's image moments, which are resistant to scaling and rotation. See https://www.youtube.com/watch?v=Nc06tlZAv_Q

Or you might try the Hausdorf distance. See http://cgm.cs.mcgill.ca/~godfried/teaching/cg-projects/98/normand/main.html