[image-capture] Missing some constraints or options

Looking at the current draft of the image capturing I can't find any
constraints or capabilities that would help fulfilling a very simple
photo-task which is taking a photo of an id-card.

Nowadays, smartphones come with a large number of cameras and choosing the
correct one is crucial for taking a photo of an object only 30cm away to
fill the sensor.

Media constraints should help selecting the correct camera, but that would
require to select a camera by supported FoV and focusDistance. At the
moment absolute information about the physical properties of a camera is
very sparse.