3. Object Recognition Using Vision
As mentioned above, the data acquisition model is only based on vision by means of a normal camera. The
goal is to extract reference or landmark objects such as doors, windows, lockers, text and figural signs, and fire
extinguishers. The objects are contextualised through a frame identifier, the type of object and position
coordinates (in the image), and they are stored in a first-in-first-out (FIFO) buffer. This buffer, also called the
“navigation history,” is later used for location estimation and navigation.
The process of object recognition starts by (a) capturing an image, followed by (b) a process that aims to
contextualise detected objects by checking if the objects are identical in previous images. Contextualisation is