ADVANCED DRIVER ASSISTANCE SYSTEM AND METHOD

• US 20200134326A1
• Filed: 12/27/2019
• Published: 04/30/2020
• Est. Priority Date: 06/28/2017
• Status: Active Grant

• Current Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Original Assignee: Huawei Technologies Co., Ltd. (Huawei Investment & Holding Co., Ltd.)
• Inventors: BOEV, Atanas, URFALIOGLU, Onay, SETIAWAN, Panji
• Granted Patent: US 11,164,012 B2
• Time in Patent Office: Days
• CPC Class Codes: B60W 2552/53 Road markings, e.g. lane marker or crosswalk
• CPC Class Codes: B60W 40/072 Curvature of the road
• CPC Class Codes: G06F 18/23 Clustering techniques
• CPC Class Codes: G06V 10/457 by analysing connectivity, e.g. edge linking, connected component analysis or slices
• CPC Class Codes: G06V 10/757 Matching configurations of points or features
• CPC Class Codes: G06V 20/588 Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road

Abstract

A driver assistance system detects lane markings in a perspective image of a road in front of a vehicle. The driver assistance system extracts a plurality of features, in particular lane markings, from the perspective image for generating a set of feature coordinates, in particular lane marking coordinates. The system generates a plurality of pairs of feature coordinates, each pair defining a straight line, and estimates a lane curvature on the basis of a subset of the pairs of feature coordinates. For each pair a straight line defined by the pair intersects a predefined target portion of the perspective image, the predefined target portion including a plurality of possible positions of a vanishing point.

Claims

1. A driver assistance system configured to detect lane markings in a perspective image of a road in front of a vehicle, the perspective image comprising a vanishing point, the driver assistance system comprising a processor configured to:

extract a plurality of features from the perspective image for generating a set of feature coordinates, wherein the set of feature coordinates defines respective positions of the plurality of features in the perspective image;

generate a plurality of pairs of feature coordinates, wherein each of the pairs of feature coordinates defines a straight line; and

estimate a lane curvature on the basis of a subset of the plurality of pairs of feature coordinates, wherein for each pair of feature coordinates of the subset of the plurality of pairs of feature coordinates, the straight line defined by the respective pair of feature coordinates intersects a predefined target portion of the perspective image, wherein the predefined target portion comprises a plurality of possible positions of the vanishing point.

2. The system of claim 1, wherein the predefined target portion has a quadrilateral shape in the perspective image.

3. The system of claim 1, wherein the processor is configured to estimate the lane curvature on the basis of the subset of the plurality of pairs of feature coordinates using a RANSAC algorithm.

4. The system of claim 3, wherein the processor is configured to estimate the lane curvature on the basis of the subset of the plurality of pairs of feature coordinates using the RANSAC algorithm by performing the following steps:

randomly selecting n pairs of feature coordinates from the plurality of pairs of feature coordinates, wherein n is a minimum number required to generate a fitting model;

generate the fitting model on the basis of the n pairs of feature coordinates;

counting the number of pairs of feature coordinates consistent with the fitting model; and

repeating the previous steps until a fitting model consistent with a large number of pairs of feature coordinates is found.

5. The system of claim 1, wherein the processor is configured to generate the plurality of pairs of feature coordinates:

by dividing the perspective image into a plurality of horizontally extending stripes, and by processing the plurality of horizontally extending stripes from bottom to top.

6. The system of claim 5, wherein for a currently processed horizontally extending stripe, of the horizontally extending stripes, the processor is configured:

to divide the currently processed horizontally extending stripe into a plurality of source windows, and to generate first feature coordinates, of a pair of feature coordinates of the plurality of pairs of feature coordinates, by selecting feature coordinates from at least one feature located in at least one of the plurality of source windows.

7. The system of claim 6, wherein for the currently processed horizontally extending stripe, the processor is configured:

to divide a neighbouring horizontally extending stripe above the currently processed horizontally stripe into a plurality of destination windows, and to generate second feature coordinates of the pair of feature coordinates by selecting feature coordinates from at least one feature located in at least one of the plurality of destination windows.

8. The system of claim 7, wherein the plurality of destination windows of the neighbouring horizontally extending stripe above the currently processed horizontally extending stripe are at least partially overlapping.

9. The system of claim 1, wherein the processor is configured:

to cluster the plurality of straight lines defined by the plurality of pairs of feature coordinates of the subset of the plurality of pairs of feature coordinates into a set of straight line clusters comprising one or more of the plurality of straight lines, and to determine for each straight line cluster a representative cluster line.

10. The system of claim 9, wherein the processor is configured to estimate the lane curvature on the basis of the set of representative cluster lines.

11. The system of claim 1, wherein the system further comprises a stereo camera configured to provide the perspective image of the road in front of the vehicle as a stereo image having a first channel and a second channel.

12. The system of claim 11, wherein the processor is configured:

to independently extract features from the first channel of the stereo image and the second channel of the stereo image, and to determine those features, which have been extracted from both the first channel and the second channel of the stereo image.

15. The system of claim 1, wherein the plurality of features are lane markings and the set of feature coordinates are lane marking coordinates.

13. A method of detecting lane markings in a perspective image of a road in front of a vehicle, the perspective image defining a vanishing point, wherein the method comprises:

extracting a plurality of features from the perspective image for generating a set of feature coordinates, wherein the set of feature coordinates defines respective positions of the plurality of features in the perspective image;

generating a plurality of pairs of feature coordinates, wherein each of the pairs of feature coordinates defines a straight line; and

estimating a lane curvature on the basis of a subset of the plurality of pairs of feature coordinates, wherein for each pair of feature coordinates of the subset of the plurality of pairs of feature coordinates, the straight line defined by the respective pair of feature coordinates intersects a predefined target portion of the perspective image, wherein the predefined target portion comprises a plurality of possible positions of the vanishing point.

14. A non-transitory computer readable medium comprising program code for performing the method of claim 13 when executed on a computer or a processor.