Abstract
Remote controlled Micro Aerial Vehicles (MAVs) have become increasingly popular in the last five years. One of the most popular of these MAVs is the Parrot AR Drone. While the predominant use of this drone is recreational, it provides a platform with a large array of functionality for testing the capabilities of MAVs. This project focused on taking a Parrot AR Drone and, using its on-board vertical camera, made it to follow a series of glyph patterns laid out along the floor that guide the AR Drone over unknown terrain. Due to computational and speed restraints of real-time control, the system made use of colour recognition to detect a point over which to hover the AR Drone, and pattern matching using glyphs for orientation. On aligning with the glyph, the AR Drone is instructed to move to the next glyph, in a series, following a predetermined route based on glyph placement. The AR Drone then traverses the series of glyphs saving incremental pictures of the ground below. To prove the capabilities of the AR Drone as a close range mapping tool, the saved images were then stitched together using panoramic imaging software to create a single image of the ground below. While the AR Drone itself is not a desirable MAV for the use of close range mapping, the results prove that this type of functionality is possible.
