Systems and methods for vehicle navigation

A navigation system for a vehicle which receives 4010, from a camera of the vehicle, a plurality of images captured from an environment of the vehicle. A first image is analysed 4020 to identify a non-semantic road feature, and a second image is analysed 4040 to identify a representation of the non-semantic road feature. Three dimensional coordinates for a point associated with the non-semantic road feature are determined 4060 based on a difference between the first and second image locations and on motion information for the vehicle between a capture of the first image and a capture of the second image. The three-dimensional coordinates for the point associated with the non-semantic road feature are sent 4070 to a server for use in updating a road navigation model. The non-semantic road feature may be the back of a sign, a building, a light pole, a road crack, or a pothole.