2-D Occupancy Grid

[bohemian]

Referring to terms which are used in the wiki section of the
Sentience, I would like to know if a "disparity (depth) map" could be
considered as a "2-D Occupancy Grid" which later can be used to build
the "3-D Occupancy Grid"?

[Bob Mottram]

An individual disparity map on its own is fairly useless, since there is so much uncertainty in individual sensor readings.  A single pixel of disparity could mean that there is an object ten metres away or thirty.  It's only as the robot moves around, updating the occupancy grid that 2D or 3D objects start to appear.