To make error correction in a position estimate of a vehicle, visual lane markings on the road can be matched with lane boundaries for the road within 3-D map data. Embodiments include obtaining a “stripe” indicative of an observed lane marking captured in a from a camera image, obtaining map data for the road on which the vehicle is located, determining the plane of the road from coordinates of lane boundaries within the map data, projecting the stripe onto the plane, and comparing the projected stripe with lane boundaries within the map data and associating the visual lane markings with the closest lane boundary. Differences between the projected stripe and the associated lane boundary can then be used for error correction in the position estimate of the vehicle.