Image based lane marking classification

Patent metadata

Field	Value
Publication number	US-10943131-B2
Application number	US-201916409035-A
Country	US
Kind code	B2
Filing date	May 10, 2019
Priority date	May 30, 2018
Publication date	Mar 9, 2021
Grant date	Mar 9, 2021

Abstract

An image processing method includes: determining a candidate track in an image of a road, wherein the candidate track is modelled as a parameterized line or curve corresponding to a candidate lane marking in the image of a road; dividing the candidate track into a plurality of cells, each cell corresponding to a segment of the candidate track; determining at least one marklet for a plurality of said cells, wherein each marklet of a cell corresponds to a line or curve connecting left and right edges of the candidate lane marking; determining at least one local feature of each of said plurality of cells based on characteristics of said marklets; determining at least one global feature of the candidate track by aggregating the local features of the plurality of cells; and determining if the candidate lane marking represents a lane marking based on the at least one global feature.

First claim

We claim:

1. An image processing method for lane classification, said method comprising:
determining a candidate track in an image of a road, wherein the candidate track is modelled as a parameterized line or curve corresponding to a candidate lane marking in the image of a road;
dividing the candidate track into a plurality of cells, each cell corresponding to a segment of the candidate track;
determining at least one marklet for a plurality of said cells, wherein each marklet of a cell corresponds to a line or curve connecting left and right edges of the candidate lane marking;
determining at least one local feature of each of said plurality of cells based on characteristics of said marklets of each respective cell by determining at least one inlier ratio of at least one of the cells as the ratio between a number of marklets determined for the cell and a number of image pixel lines comprised by the cell;
determining at least one global feature of the candidate track by aggregating the local features of the plurality of cells;
determining if the candidate lane marking represents a lane marking based on the at least one global feature;
classifying, if it is determined that the candidate lane marking represents a lane marking, the lane marking as a solid line or as a dashed based on the local features of a plurality of said cells including the at least one inlier ratio.

2. The image processing method of claim 1 , wherein each marklet of a cell is defined by a horizontal row of pixels of said image of a road connecting left and right edges of the candidate lane marking extending in the vertical direction in the image of a road.

3. The image processing method of claim 1 , wherein said determining at least one local feature includes determining an optical feature of a marklet of a cell by determining the gradient pattern of gray values of a line or curve of pixels of the candidate lane marking corresponding to the marklet.

4. The image processing method of claim 1 , wherein said determining at least one local feature includes determining a geometric feature of a marklet of a cell by determining the position of the marklet relative to the candidate track.

5. The image processing method of claim 1 , wherein the candidate track is divided into a plurality of cells such that each cell corresponds to a segment of the candidate track having a fixed length, wherein the length of each segment corresponds to a fixed length as measured on the road surface.

6. The image processing method of claim 5 , wherein the fixed length of each cell is in the range 10 cm to 5 m.

7. The image processing method of claim 1 , wherein the lane marking is classified as a solid line or as a dashed line based on the variance of the inlier ratio for consecutive cells of the candidate track.

8. The image processing method of claim 1 , wherein the lane marking is classified as a solid line or as a dashed line based on the variance of the inlier ratio of cells captured in consecutive images of the road for cells positioned in the near range of the respective camera used for capturing the images of the road.

9. The image processing method of claim 7 , wherein the lane marking is classified as a solid line or as a dashed line based on the variance of the inlier ratio of cells, such that a line marking classified as a solid line has a smaller variance of inlier ratio of cells than that of a line marking classified as a dashed line.

10. The image processing method of claim 1 , wherein if it is determined that the candidate lane marking represents a lane marking, a range of the lane marking is determined based on the local features of a plurality of said cells, wherein the range of the lane marking corresponds to the non-obstructed length of the lane marking.

11. The image processing method of claim 10 , wherein the range of the lane marking is determined by comparing the at least one local feature of a plurality of cells to a cell model to determine which of the cells match a lane marking segment.

12. The image processing method of claim 11 , wherein if it is determined that a cell matches a lane marking segment, the length of the segment of the candidate track corresponding to the cell is added to the range of the lane marking, and the at least one local feature of the cell is used for updating the cell model.

13. The image processing method of claim 1 , wherein determining if the candidate lane marking represents a lane marking based on the at least one global feature includes performing a classification using a machine learning classifier, an artificial neural network, or a support vector machine.

14. A system comprising:
means for determining a candidate track in an image of a road, wherein the candidate track is modelled as a parameterized line or curve corresponding to a candidate lane marking in the image of a road;
means for dividing the candidate track into a plurality of cells, each cell corresponding to a segment of the candidate track;
means for determining at least one marklet for a plurality of said cells, wherein each marklet of a cell corresponds to a line or curve connecting left and right edges of the candidate lane marking;
means for determining at least one local feature of each of said plurality of cells based on characteristics of said marklets of each respective cell including means for determining at least one inlier ratio of at least one of the cells as the ratio between a number of marklets determined for the cell and a number of image pixel lines comprised by the cell;
means for determining at least one global feature of the candidate track by aggregating the local features of the plurality of cells; and
means for determining if the candidate lane marking represents a lane marking based on the at least one global feature,
wherein if it is determined that the candidate lane marking represents a lane marking, further comprising means for classifying the lane marking as a solid line or as a dashed based on the local features of a plurality of said cells including the at least one inlier ratio.

15. A system comprising a processor configured to:
determine a candidate track in an image of a road, wherein the candidate track is modelled as a parameterized line or curve corresponding to a candidate lane marking in the image of a road;
divide the candidate track into a plurality of cells, each cell corresponding to a segment of the candidate track;
determine at least one marklet for a plurality of said cells, wherein each marklet of a cell corresponds to a line or curve connecting left and right edges of the candidate lane marking;
determine at least one local feature of each of said plurality of cells based on characteristics of said marklets of each respective cell by determining at least one inlier ratio of at least one of the cells as the ratio between a number of marklets determined for the cell and a number of image pixel lines comprised by the cell;
determine at least one global feature of the candidate track by aggregating the local features of the plurality of cells;
determine if the candidate lane marking represents a lane marking based on the at least one global feature; and
classify, if it is determined that the candidate lane marking represents a lane marking, the lane marking as a solid line or as a dashed based on the local features of a plurality of said cells including the at least one inlier ratio.

16. The system of claim 15 , wherein each marklet of a cell is defined by a horizontal row of pixels of said image of a road connecting left and right edges of the candidate lane marking extendin

Assignees

• Aptiv Tech Ltd

Inventors

• Su Yu
• Paus Andre
• Zhao Kun
• Meuter Mirko
• Nunn Christian

Classifications

• G06F18/24

  Classification techniques · CPC title

• G06V20/588 Primary

  Recognition of the road, e.g. of lane markings; Recognition of the vehicle driving pattern in relation to the road · CPC title

• G06T2207/30256

  Lane; Road marking · CPC title

• B60W30/12

  Lane keeping · CPC title

• G06K9/00798 Primary

  Physics · mapped topic