Given a scene consisting of one or more objects from a collection of known objects O1,...,On, and an image of that scene taken from an unknown position and orientation, determine which of the O1,...,On are present in the scene. We need to determine the position and orientation (pose) of each object.
The objects are represented as polyhedral approximations (solids, no curves). This is a very general representation, but leads to lengthy computation when accuracy is required. Another approach is to use generalized cylinders.