hey hey community,

i am planning an interactive white model installation.
that means that there are several (about 20) white cubes distributes among 10x6 meters. a projector and a camera are covering the whole area from 8 meters from above.

i want the camera to track people
and the projector to display variable images on the top surface of the white cubes.
ultimately there should be connecting lines from and to every active surface.

so, i figured i need the OpenCV library, and search for two kinds of blobs:

1. the white rectangular surfaces of the cubes ('screens')
2. the shadows of people. ('shadows-blobs')

i want to know the size and position of the surfaces ('screens').

when people, ('shadows-blobs') are approaching the 'screens',
i want to be abled to display .jpg images on them.

when screens are active i want to draw lines between them, ultimately in a very complex network.

///
this idea is the ultimate goal. it is variable - but i could imagine defining the 'screens' by 'hand' - how would i do that?

i hope you can give me some tips how to approach these problems.
is this idea actually realistic?
what do you think?

best regards

TIL