The system comprises a Training Module for creating object libraries. These are formed by keypoints drawn
from a set of images with different views of each object. Each keypoint is characterised by its neighbourhood
by means of features extracted by the OpenSURF library [12]. After training, these libraries are then used to
recognise the objects in the images captured by the camera (see Section 3 or [5]). The system is controlled by a
Main Module for real-time localisation and navigation.