How to calculate distance between 2 overlaped images edge to edge?

I reconsruct 3D model from images. Now I calculate to error between reconstruction image and original image. How to calculate this error. Please help me.

Answer 1

[Image Analyst]

It's not clear exactly where that line is. It looks arbitrary. I suggest you go with the figures of merit suggested by Mohammad instead. You could also calculate the area of the dark gray mismatch if you want, though it's a bit redundant given the other information.

[Image Analyst]

Regarding your explanation in the comment to Mohammad, just XOR the two images, then run bwdist() on the result and take the max.

% Do an exclusive OR to find mismatches.
xorImage = xor(bw1, bw2);
% Get the Euclidean Distance Transform.
edtImage = bwdist(xorImage);
% Find max and double it because EDT max is the radius, not the diameter.
maxDistance = 2 * max(edtImage(:));

[Image Analyst]

This misalignment looks so small that the distance transform to get the width of the lines doesn't make much sense. Why don't you just sum the xor image?

% Do an exclusive OR to find mismatches.
xorImage = xor(bw1, bw2);
% Count the different pixels
numDiffPixels = sum(xorImage(:));

[Jagaa]

I need to calculate 2 types error:

1. Sum the xor image - I use this error. I think this is sum area error.
2. Image accuracy error. Max, min distance. I didn't solve this type error.

[Image Analyst]

I gave you code for all of that. For #2 in my first comment, and for #1 in the last comment. Of course the min error is 0 pixels, or 1 pixel if you don't count 0 as being an error.

[Jagaa]

Thank you very much. I understand what should I do.