Lane detection

from Wikipedia, the free encyclopedia

Lane recognition refers to the automated identification of one's own lane and the neighboring lanes of road vehicles . On the basis of environmental features such as lane markings, information is obtained that is further processed in driver assistance systems or systems of autonomous vehicles .

In lane detection, an image of the surroundings is first obtained via sensors , from which an attempt is then made using an algorithm (often using image processing means ) to calculate one's own lane and possibly also neighboring lanes. Globally localizing systems such as GPS are not primarily used here, as they are not sufficiently accurate, require an accurate, known map of the area and (e.g. in tunnels) work unreliably. Camera-based systems require a clear picture of the road surface. 2D systems do not provide any distance information and are therefore unsuitable as sensors for other assistance systems, such as emergency brake assistants . 3D cameras are able to provide spatial information. Lidar systems are in the experimental stage.

The first attempts at lane recognition took place in the 1980s (see Prometheus (research program) ). Since the beginning of the 21st century, some vehicles have been equipped as standard with assistance systems such as lane departure warning systems, which can recognize lanes among other things.

Web links