Detecting changes between two pictures

Question

I am a beginner in python & image processing.I have the following image.

I would like to detect the changes between this picture and another one where the other one may be

• taken from a slightly different angle or the object is translated
• have different light conditions

and the changes may be

• a change in color in part of the object
• an extra or missing part

After various searches I thought about using ORB to detect the matching parts and remove them from the picture then use contour to extract and compare the difference.

But I cannot seem to find a way to remove the matching parts from the two images.

I am open for all suggestions / better way to approach the problem.

---

edit: Sorry , forgot to mention that the color change could either be white or pink

Answer 1

Looking at your image, it appears there are three dominant colors. If this is always the case the first thing that comes to mind is to apply a color K mean algorithm with three clusters, like explained here.

The center color of each cluster would then give you information on the color of the tubes, and the size of each cluster (# of pixels belonging to that cluster) would give you if there are extra or missing parts.