The US Department of Defense has issued a challenge for computer vision to correctly classify the state of a building after disasters:

Your model must predict an output PNG image with height and width of 1024 pixels, where each pixel value corresponds to the predicted class at that place in the input image:

0: no building
1: undamaged building
2: building with minor damage
3: building with major damage
4: destroyed building

This makes for very interesting games played at the superficial level with paint or loose objects.

For example could computer vision distinguish easily the sort of building under perpetual construction (as often is the case in developing nations) from one that has been damaged? Or in a similar vein, if buildings are maintained poorly in markets that lack readily available materials, what constitutes minor damage?