Road edge extraction method and device

Abstract

The invention discloses a road edge extraction method, and relates to the technical field of unmanned driving, and the method comprises the steps: obtaining the original point cloud data of a road environment; preprocessing the single-frame point cloud data; screening out point cloud data obtained by scanning the scanning line which scans the road edge and has a distance with the vehicle smaller than a threshold value; establishing a rectangular coordinate system by taking the forward direction of the vehicle as an x axis, screening out points of which the abscissa is greater than 0, dividing the points into right side road edge candidate points and left side road edge candidate points according to whether the ordinate is greater than 0, and sorting the screened points from small to large according to direction angles; road edge feature points are screened through the vertical height difference and the horizontal distance threshold value; converting the road edge feature points into a map coordinate system; and according to the left-side road edge points and the right-side road edge points extracted from the plurality of continuous frames of point cloud data, generating a road edge line of a corresponding side. According to the method, the influence of obstacles on road edge point extraction can be effectively avoided, and the generated road edge line is high in precision.

Description

technical field [0001] The present invention relates to the technical field of unmanned driving, in particular to a roadside extraction method and device. Background technique [0002] For unmanned driving systems and advanced assisted driving systems, the environmental perception system is the interaction port between the intelligent vehicle platform and the surrounding traffic scenes, and is also the front-end input of the intelligent vehicle platform's motion decision-making and planning control system. Road edge recognition is an important and serious content in environment perception system. [0003] Existing roadside extraction methods deal with obstacles blocking the roadside, such as vehicles on both sides of the collection vehicle or pedestrians blocking the roadside, the extracted roadside points may fall on vehicles or pedestrians, resulting in the extracted roadside The along point is not accurate. In addition, when extracting the curved curb at the traffic int...

AI Technical Summary

Problems solved by technology

[0003] Existing roadside extraction methods deal with obstacles blocking the roadside, such as vehicles on both sides of the collection vehicle or pedestrians blocking the roadside, the extracted roadside points may fall on vehicles or pedestrians, resulting in the extracted roadside Inaccurate along the point

In addition, when extracting the curved curb at the traffic intersection, the extracted curb points may fall outside the curb

In addition, when the collected vehicle changes lanes, the angle between the vehicle’s forward direction and the roadside is too large, so the extracted roadside points will also fall outside the roadside.

Images