Lane Departure Warning

Real-time lane detection and tracking in embedded systems Lane detection and tracking using a monocular sensor connected to an embedded processor. The system is able to determine the position of the vehicle inside its own lane, alert the driver in case of lane departures (notice that the lane is painted in orange when the wheels get closer to the detected lane markings). The system has been tested in a wide range of conditions, including night, rain, traffic jams, etc.

This video shows a lane detection, tracking and modeling approach that is able to model multiple lanes (ego-lane and lateral lanes), estimate their existence confidence, and other parameters, such as their curvature. The view distance of the lanes is also estimated according to the available lane marking measurements.