We're a group of mechanical engineering students trying to clean up a piece of honeycomb material, after it's been cut on a CNC router. We're not looking for any specific answer to a coding problem but we're wondering what might be a good way to identify burrs from a picture.

So far, our idea is to take a picture at an angle perpendicular to the material and filter out the background, in a way so that we can see through the honeycomb. After editing the picture we would get a contrast, where the honeycomb pattern turns white and everything else turns black. The main problem is then to identify the burrs (they would presumably also be white) as deviations from the honeycomb pattern.

The solution we imagine at the moment involves a neural network detecting deviations from the pattern, but we're wondering if any of you might know of an easier solution. If not, it would be a great help to know what coding languages or programs might have the capabilities to do some of the stuff described above. Any advice is helpful - thanks in advance.

Honeycomb normal:

Honeycomb contrast: