Diseño e implementación de un prototipo de detección de carril mediante visión artificial

Abstract

A system to detect the lane lines that are clearly marked on the roads by using artificial vision, is implemented in this work, it mainly consists in a Raspberry Pi 2 Model B, a screen and a camera. This system can be installed in a vehicle and based on the lane recognition, it detects lane departure and warns the driver through visual and audible alerts. The system does not distinguish between lane departures performed voluntarily or involuntarily. The prototype displays in the screen the distance in centimeters, between the vehicle’s side edges and the closest lane’s line, it also displays the lane width in meters. The system performance depends on the proper camera installation inside the vehicle, because several parameters to compute the above distances depend on its location. The module works when there is an adequate lighting level and do not exist weather conditions that reduce the camera visibility, like the presence of fog, rain, or excessive illumination.