Low-cost visually servoed tracked vehicle

摘要

There are many applications employing image processing and/or control algorithms. Video surveillance, visual inspection of manufactured goods are among the examples. Self - driven cars are today one of the hottest topic. Using powerful image processing, control and learning technologies, self-driven vehicles appears to be just around the corner. Another possible application are robots in charge of moving manufactured parts between the stations. Indeed, robotic-carriers (wheeled or tracked) may be visually guided to follow colored lines painted on the floor in order to automatically transport parts between the stations. This paper presents the construction and the implementation of such a visually guided vehicle which is able to follow a white line on a dark background. The vehicle is controlled by a so called embedded computer (a Raspberry Pi board) which uses a camera, minimal image processing and simple control law in order to command the vehicle to “stay on the line”. A description of the vehicle, the setup of the Raspberry Pi board and the image processing and control are also described. The obtained results are also presented.