Method and system for lane departure warning

• US 8,902,053 B2
• Filed: 03/20/2012
• Issued: 12/02/2014
• Est. Priority Date: 11/16/2011
• Status: Active Grant

• Current Assignee: Industrial Technology Research Institute
• Original Assignee: Industrial Technology Research Institute
• Inventors: Wu, Yi-Ta
• Primary Examiner(s): Wu, Daniel
• Assistant Examiner(s): OTT, FREDERICK R
• Application Number:
  US13/424,378
• Publication Number:
  US 20130120125A1
• Time in Patent Office: 987 Days
• Field of Search: 340/439, 340/437, 340/576
• US Class Current: 340/438
• 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
• CPC Class Codes: G08G 1/167 Driving aids for lane monitoring, lane changing, e.g. blind spot detection

Abstract

A method and a system for lane departure warning are provided. The method is as follows. An original image is segmented into a plurality of regional images. Next, characteristics of each regional image are analyzed, and accordingly non-lane line regions are removed from the regional images, so as to obtain a plurality of lane line candidates. Then, a plurality of lane lines are determined from the lane line candidates according to a location of each lane line candidate in the original image. Finally, the lane lines are distinguished into left lane lines and right lane lines, and a variation of an angle between each left lane line and a horizontal line and a variation of an angle between each right lane line and the horizontal line are analyzed, so as to judge whether a vehicle departs from a lane and send a lane departure warning.

Claims

1. A method for lane departure warning, applicable to an electronic device, the method comprising:

segmenting an original image into a plurality of regional images;

analyzing a plurality of characteristics of each of the regional images, and accordingly removing at least one non-lane line region from the regional images, so as to obtain a plurality of lane line candidates;

determining a plurality of lane lines from the lane line candidates according to a location of each of the lane line candidates in the original image;

distinguishing the lane lines into at least one left lane line and at least one right lane line; and

analyzing a variation of an angle between each of the at least one left lane line and the at least one right lane line and a horizontal line, so as to judge whether the electronic device departs from a lane and send a lane departure warning, wherein the step of analyzing the variation of the angle between each of the at least one left lane line and the at least one right lane line and the horizontal line, so as to judge whether the electronic device departs from the lane comprises;

comparing the angle of each of the at least one left lane line and the angle of each of the at least one right lane line with a first angle threshold and a second angle threshold respectively;

when the angle of the left lane line is greater than the first angle threshold and the angle of the right lane line is smaller than the second angle threshold, determining that departure to the left occurs; and

when the angle of the right lane line is greater than the first angle threshold and the angle of the left lane line is smaller than the second angle threshold, determining that departure to the right occurs.

2. The method for lane departure warning according to claim 1, wherein the step of segmenting the original image into the plurality of regional images comprises:

calculating a histogram illustrating a distribution of pixel values of a plurality of pixels in the original image; and

performing a grouping process on the histogram, and restoring a result of the grouping process to the original image, so as to obtain the regional images.

3. The method for lane departure warning according to claim 1, wherein the step of analyzing the characteristics of each of the regional images, and accordingly removing the at least one non-lane line region from the regional images, so as to obtain the lane line candidates comprises:

detecting a plurality of lane markers in each of the regional images, and classifying the lane markers into a strong edge, a soft edge, or a non-edge according to a gradient value of each of the lane markers; and

analyzing at least one characteristic of the lane markers classified as the strong edge or the soft edge, and accordingly removing the at least one non-lane line region from the lane markers, so as to obtain the lane line candidates.

4. The method for lane departure warning according to claim 3, wherein the step of classifying the lane markers as the strong edge, the soft edge, or the non-edge according to the gradient value of each of the lane markers comprises:

using a plurality of different gradient thresholds to distinguish the lane markers into the strong edge, the soft edge, or the non-edge for the lane markers of different categories.

5. The method for lane departure warning according to claim 3, wherein the step of analyzing the at least one characteristic of the lane markers classified as the strong edge or the soft edge, and accordingly removing the at least one non-lane line region from the lane markers, so as to obtain the lane line candidates comprises:

calculating a first ratio of a maximum value to a minimum value in the width and the height of each of the lane markers;

calculating a second ratio of a pixel number of each of the lane markers to the width and the height;

judging whether the first ratio is greater than a first threshold and whether the second ratio is greater than a second threshold;

if yes, determining that the lane marker is a lane line candidate; and

if not, deteiiiiining that the lane marker is a non-lane line region.

6. The method for lane departure warning according to claim 5, wherein the step of analyzing the characteristics of the lane markers classified as the strong edge or the soft edge, and accordingly removing the at least one non-lane line region from the lane markers, so as to obtain the lane line candidates further comprises:

using connected component labeling to label the lane markers that are connected as the same lane marker.

7. The method for lane departure warning according to claim 1, wherein the step of determining the lane lines from the lane line candidates according to the location of each of the lane line candidates in the original image comprises:

using a chain-code to detect a contour of each of the lane line candidates;

analyzing the contours of the lane line candidates to find a plurality of straight line regions; and

removing at least one non-lane region from the straight line regions according to locations of the straight line regions in the original image, so as to obtain the lane lines.

8. The method for lane departure warning according to claim 1, wherein the step of distinguishing the lane lines into the at least one left lane line and the at least one right lane line comprises:

distinguishing each of the lane line regions by using a midpoint of a base of the original image as a reference, wherein when an intersection point of the lane line region and the base is on the left of the midpoint, the lane line region is determined as a left lane line; and

when the intersection point of the lane line region and the base is on the right of the midpoint, the lane line region is determined as a right lane line.

9. The method for lane departure warning according to claim 1, wherein lane lines for judging whether the electronic device departs from the lane are a left lane line and a right lane line closest to a midline of the original image, and the angle of the left lane line and the angle of the right lane line are acute angles.

10. A system for lane departure warning, comprising:

an image segmentation module, for segmenting an original image into a plurality of regional images;

a lane line region detection module, for analyzing a plurality of characteristics of each of the regional images, and accordingly removing at least one non-lane line region from the regional images, so as to obtain a plurality of lane line candidates;

a lane line determination module, for determining a plurality of lane lines from the lane line candidates according to a location of each of the lane line candidates in the original image;

a lane departure judgment module, for distinguishing the lane lines into at least one left lane line and at least one right lane line, and analyzing a variation of an angle between each of the at least one left lane line and the at least one right lane line and a horizontal line, so as to judge whether departure from a lane occurs, wherein the lane departure judgment module comprises comparing the angle of each of the at least one left lane line and the angle of each of the at least one right lane line with a first angle threshold and a second angle threshold respectively, wherein when the angle of the left lane line is greater than the first angle threshold and the angle of the right lane line is smaller than the second angle threshold, it is determined that departure to the left occurs; and

when the angle of the right lane line is greater than the first angle threshold and the angle of the left lane line is smaller than the second angle threshold, it is determined that departure to the right occurs; and

a warning module, for sending a lane departure warning when the lane departure judgment module judges that departure from the lane occurs.

11. The system for lane departure warning according to claim 10, wherein the image segmentation module comprises calculating a histogram illustrating a distribution of pixel values of a plurality of pixels in the original image;

and performing a grouping process on the histogram, and restoring a result of the grouping process to the original image, so as to obtain the regional images.

12. The system for lane departure warning according to claim 10, wherein the lane line region detection module comprises:

a lane marker detection unit, for detecting a plurality of lane markers in each of the regional images, and classifying the lane markers into a strong edge, a soft edge, or a non-edge according to a gradient value of each of the lane markers; and

a lane marker filtering unit, for analyzing at least one characteristic of the lane markers classified as the strong edge or the soft edge, and accordingly removing the at least one non-lane line region from the lane markers, so as to obtain the lane line candidates.

13. The system for lane departure warning according to claim 12, wherein the lane marker detection unit further comprises using a plurality of different gradient thresholds to distinguish the lane markers into the strong edge, the soft edge, or the non-edge for the lane markers of different categories.

14. The system for lane departure warning according to claim 12, wherein the lane marker filtering unit comprises:

calculating a first ratio of a maximum value to a minimum value in the width and the height of each of the lane markers;

calculating a second ratio of a pixel number of each of the lane markers to the width and the height;

judging whether the first ratio is greater than a first threshold and whether the second ratio is greater than a second threshold;

if yes, detennining that the lane marker is a lane line candidate; and

if not, determining that the lane marker is a non-lane line region.

15. The system for lane departure warning according to claim 14, wherein the lane marker filtering unit futher comprises using connected component labeling to label the lane markers that are connected as the same lane marker.

16. The system for lane departure warning according to claim 10, wherein the lane line determination module comprises using a chain-code to detect a contour of each of the lane line candidates, analyzing the contours of the lane line candidates to find a plurality of straight line regions, and removing at least one non-lane region from the straight line regions according to locations of the straight line regions in the original image, so as to obtain the lane lines.

17. The system for lane departure warning according to claim 10, wherein the lane departure judgment module comprises distinguishing each of the lane line regions by using a midpoint of a base of the original image as a reference, wherein when an intersection point of the lane line region and the base is on the left of the midpoint, the lane line region is determined as a left lane line;

and when the intersection point of the lane line region and the base is on the right of the midpoint, the lane line region is determined as a right lane line.

18. The system for lane departure warning according to claim 10, wherein lane lines for judging whether departure from the lane occurs are a left lane line and a right lane line closest to a midline of the original image, and the angle of the left lane line and the angle of the right lane line are acute angles.