DETECTION OF VEHICLES IN IMAGES OF A NIGHT TIME SCENE

• US 20110164789A1
• Filed: 07/13/2009
• Published: 07/07/2011
• Est. Priority Date: 07/14/2008
• Status: Active Grant

• Current Assignee: National ICT Australia Limited
• Original Assignee: National ICT Australia Limited
• Inventors: Robert, Kosita
• Granted Patent: US 8,798,314 B2
• Time in Patent Office: Days
• US Class Current: 382/104
• CPC Class Codes: G06V 10/44 Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
• CPC Class Codes: G06V 20/54 of traffic, e.g. cars on the road, trains or boats
• 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

The invention concerns the detection of vehicles in images of a night time scene. In particular, but not limited to, the invention concerns a traffic surveillance system that is used to detect and track vehicles to determine information about the detected and tracked vehicles. Candidate pair of headlights are identified 900 in an image based on luminance of points in the image. These candidates are then verified 902 by identifying 400i a sub-image of the image sized to include a candidate vehicle having the pair of candidate headlights; and determining whether the candidate vehicle is a vehicle represented in the image by testing 400k the sub-image for the presence of predetermined features of a vehicle other than the headlights. Aspects of the invention include a method, software and computer hardware.

Claims

1. A method of detecting a vehicle represented in an image of a night time scene, the method comprising the steps of:

(a) identifying a candidate pair of headlights in an image based on luminance of points in the image;

(b) identifying a sub-image of the image sized to include a candidate vehicle having the pair of candidate headlights; and

(c) determining whether the candidate vehicle is a vehicle represented in the image by testing the sub-image for the presence of predetermined features of a vehicle other than the headlights.

2. The method according to claim 1, wherein the step of identifying a pair of candidate headlights comprises identifying a mass of neighboring points in the image as including a candidate headlight such that the points of the mass have predetermined luminance values.

5. The method according to claim 2, wherein identifying a pair of candidate headlights comprises estimating a location of a candidate headlight in a mass.

9. The method according to claim 2, wherein if the mass is larger than a predetermined size, the method further comprises the step of estimating the location of two candidate headlights to form a pair of candidate headlights, the two locations being the most separated locations within the mass that are laterally aligned.

10. The method according to claim 1 wherein the sub-image is sized and shaped to substantially include only a representation of the candidate vehicle.

11. The method according to claim 1, wherein the size of the sub-image is based on the position and length between the pair of candidate headlights.

12. The method according to claim 1, wherein the method further comprise the step of scaling the sub-image to a predetermined size.

13. The method according to claim 1, wherein the predetermined feature of step (c) includes at least one part of the vehicle, such as a windscreen.

15. The method according to claim 1, wherein testing the sub-image for the presence of the predetermined feature comprises determining whether the predetermined feature is a predetermined spatial relationship with the candidate headlights.

16. The method according to claim 1, wherein the step of testing the sub-image for the presence of the predetermined feature comprises applying a mask to the sub-image that outlines the predetermined feature to determine whether points having a predetermined characteristic are located in the sub-image within the outline of the predetermined feature.

17. The method according to claim 1, wherein the predetermined feature is the appearance of the vehicle in the sub-image.

18. The method according to claim 1, wherein step (c) comprises:

determining one or more descriptors to represent a sub-image; and

testing the one or more descriptors on a trained classifier to determine whether the candidate vehicle of the sub-image is a vehicle represented in the image.

19. The method according to claim 1, wherein the image of the night time scene is one image in a time ordered sequence of images that the method is performed, and the method further comprises matching the detected vehicle in subsequent images of the sequence.

20. The method according to claim 1, wherein the method further comprises the step of classifying the detected vehicle.

21. The method according to claim 20, wherein classifying the detected vehicle comprises determining the distance between the headlights of the identified pair of headlights of the detected vehicle.

22. The method according to claim 20, wherein classifying the detected vehicle comprises:

determining one or more descriptors to represent a sub-image; and

testing the one or more descriptors on a trained classifier to classify the vehicle using machine learning techniques.

23. The method according to claim 1, wherein the method further comprises the initial step of identifying lanes in the image, and

performing the method on the area of the image bounded by a vertical distance in the image above the identified lane.

24. The method according to claim 1, wherein the method further comprises the step of extracting traffic information about the detected vehicle from the image.

25. The method according to claim 1, wherein the method comprises the step of translating all points in an image from image coordinate values to a real world values.

3-4. -4. (canceled)

6-8. -8. (canceled)

14. (canceled)

26. (canceled)

27. A computer for detecting a vehicle represented in an image of a night time scene, the computer comprising:

input means to receive the image of the night time scene;

a processor to (a) identify a candidate pair of headlights in the image based on luminance of points in the image, (b) identify a sub-image of the image sized to include a candidate vehicle having the pair of candidate headlights, and (c) to determine whether the candidate vehicle is a vehicle represented in the image by testing the sub-image for the presence of predetermined features of a vehicle other than the headlights; and

output means to indicate whether a vehicle has been detected in the image.