Fast license plate verifier

• US 8,229,168 B2
• Filed: 02/20/2008
• Issued: 07/24/2012
• Est. Priority Date: 02/20/2008
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Geva, Amir, Simonsson, Rutger, Stromback, Jan Henrik, Walach, Eugeniusz
• Primary Examiner(s): Wu, Jingge
• Assistant Examiner(s): YENTRAPATI, AVINASH
• Application Number:
  US12/033,899
• Publication Number:
  US 20090208059A1
• Time in Patent Office: 1,616 Days
• Field of Search: 382/100, 382/104, 382/105, 382/173, 382/181
• US Class Current: 382/104
• CPC Class Codes: G06V 20/625 License plates
• CPC Class Codes: G06V 20/63 Scene text, e.g. street names
• CPC Class Codes: G08G 1/017 identifying vehicles G08G1/015, G08G1/054 take precedence
• CPC Class Codes: G08G 1/14 indicating individual free spaces in parking areas

Abstract

A method for identifying vehicles, including capturing a first image of a first vehicle using a first camera at a first position and a second image of the first vehicle using a second camera at a second position different from the first position. The method further includes determining a transformation between the first image and the second image. A third image of a second vehicle using the first camera is captured, and the transformation is applied to the third image to generate a fourth image of the second vehicle. The fourth image is analyzed, using a database of identified vehicles, to determine an identity of the second vehicle.

Claims

1. A method for identifying vehicles, comprising:

capturing a first image of a first vehicle using a first camera at a first position and a second image of the first vehicle using a second camera at a second position different from the first position;

determining a transformation between the first image and the second image;

capturing a third image of a second vehicle using the first camera;

applying the transformation to the third image to generate a fourth image of the second vehicle; and

analyzing the fourth image using a database of identified vehicles to determine an identity of the second vehicle;

capturing a fifth image of a third vehicle using the first camera and a sixth image of the third vehicle using the second camera;

determining an alternate transformation between the fifth image and the sixth image; and

determining a mean transformation as an average of the transformation and the alternate transformation, and wherein applying the transformation comprises applying the mean transformation;

capturing a seventh image of a fourth vehicle using the first camera and an eighth image of the fourth vehicle using the second camera;

determining a further transformation between the seventh image and the eighth image; and

determining if the further transformation exceeds the mean transformation by a preset threshold;

and comprising, in response to a positive determination that the further transformation exceeds the mean transformation by the preset threshold, identifying a change in an imaging parameter of one camera chosen from the first camera and the second camera.

2. The method according to claim 1, wherein capturing the first image and the second image comprises identifying a license plate of the first vehicle in the first image and in the second image, and in response to identifying the license plate normalizing the first and second images to produce normalized first and second images, and wherein determining the transformation comprises determining the transformation in response to the normalized first and second images.

3. The method according to claim 1, wherein capturing the third image of the second vehicle comprises identifying a license plate in the third image and in response thereto performing a positive identification of the second vehicle.

4. The method according to claim 1, wherein capturing the first image using the first camera comprises providing license plate identification criteria to the first camera, and adaptively changing the criteria in response to an analysis of the first image.

5. The method according to claim 1, wherein capturing the third image comprises determining that the identity of the second vehicle cannot be found from the third image.

6. The method according to claim 1, wherein analyzing the fourth image comprises:

identifying a license plate in the fourth image;

recognizing at least some alphanumeric characters in the license plate; and

matching the at least some alphanumeric characters with license plate characters of known vehicles stored in the database to determine the identity.

7. The method according to claim 1, and comprising:

capturing a ninth image of the second vehicle using the second camera;

identifying the second vehicle from the ninth image; and

storing the ninth image in the database,

and wherein analyzing the fourth image comprises using the database to compare the fourth image with the ninth image so as to determine the identity of the second vehicle.

8. Apparatus for identifying vehicles, comprising:

a first camera which is located at a first position and which is configured to capture a first image of a first vehicle and to capture a third image of a second vehicle;

a second camera which is located at a second position different from the first position and which is configured to capture a second image of the first vehicle; and

a processing unit which is configured to determine a transformation between the first image and the second image, to apply the transformation to the third image to generate a fourth image of the second vehicle, and to analyze the fourth image using a database of identified vehicles to determine an identity of the second vehicle;

capturing a fifth image of a third vehicle using the first camera and a sixth image of the third vehicle using the second camera;

determining an alternate transformation between the fifth image and the sixth image; and

determining a mean transformation as an average of the transformation and the alternate transformation, and wherein applying the transformation comprises applying the mean transformation;

capturing a seventh image of a fourth vehicle using the first camera and an eighth image of the fourth vehicle using the second camera;

determining a further transformation between the seventh image and the eighth image; and

determining if the further transformation exceeds the mean transformation by a preset threshold;

and comprising, in response to a positive determination that the further transformation exceeds the mean transformation by the preset threshold, identifying a change in an imaging parameter of one camera chosen from the first camera and the second camera.

9. The apparatus according to claim 8, wherein capturing the first image and the second image comprises identifying a license plate of the first vehicle in the first image and in the second image, and in response to identifying the license plate normalizing the first and second images to produce normalized first and second images, and wherein determining the transformation comprises determining the transformation in response to the normalized first and second images.

10. The apparatus according to claim 8, wherein the processing unit is configured to apply license plate identification criteria including parallelism of edges to the first image, and in response thereto to identify a license plate of the first vehicle and perform a positive identification of the first vehicle.

11. The apparatus according to claim 8, wherein capturing the third image comprises determining that the identity of the second vehicle cannot be found from the third image.

12. The apparatus according to claim 8, wherein analyzing the fourth image comprises:

identifying a license plate in the fourth image;

recognizing at least some alphanumeric characters in the license plate; and

matching the at least some alphanumeric characters with license plate characters of known vehicles stored in the database to determine the identity.

13. The apparatus according to claim 8, and comprising:

capturing a ninth image of the second vehicle using the second camera;

identifying the second vehicle from the ninth image; and

storing the ninth image in the database,

and wherein analyzing the fourth image comprises using the database to compare the fourth image with the ninth image so as to determine the identity of the second vehicle.