Methods and systems for verifying automatic license plate recognition results

• US 8,483,440 B2
• Filed: 04/13/2011
• Issued: 07/09/2013
• Est. Priority Date: 04/13/2011
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Fan, Zhigang, Kozitsky, Vladimir, Burry, Aaron Michael
• Primary Examiner(s): AZARIAN, SEYED H
• Application Number:
  US13/086,152
• Publication Number:
  US 20120263352A1
• Time in Patent Office: 818 Days
• Field of Search: 382/100, 382103-107, 382/155, 382/162, 382/168, 382/173, 382/181, 382/189, 382/194, 382/199, 382/209, 382/224, 382/232, 382/254, 382/274, 382/276, 382286-292, 382/305, 382/312, 340/425.5, 340/539.11, 340/931, 235/384
• US Class Current: 382/105
• CPC Class Codes: G06V 10/255 Detecting or recognising potential candidate objects based on visual cues, e.g. shapes
• CPC Class Codes: G06V 20/63 Scene text, e.g. street names

Abstract

Methods and systems for verifying automatic license plate recognition results providing an input image of a license plate to a processing system that extracts bitmap data from the provided input image and determines a particular license template image associated with the license plate input image. The processing system segments the input image bitmap data into text region bitmap data and outer region bitmap data and then matches the input image text region bitmap data and input image outer region bitmap data against text region bitmap data and outer region bitmap data, respectively, of the license plate image template to determine if there is a match between the license plate of the input image and the template license plate.

Claims

1. A method for verifying automatic license plate recognition results, the method comprising:

receiving, in a processing system, an input image of a license plate;

determining, in said processing system, a particular license template image associated with said license plate input image by identifying four corners of a text region of said license plate input image, identifying a plate layout based on said identified four corners, and determining said particular license plate template image based on said plate layout;

extracting, in said processing system, bitmap data from said provided input image;

segmenting, in said processing system, said bitmap data into text region bitmap data and outer region bitmap data;

matching, in said processing system, said text region bitmap data against a text region bitmap of said license plate image template; and

matching, in said processing system, said outer region bitmap data against outer region bitmap of said template.

2. The method of claim 1, further comprising normalizing said license plate input image and said license plate template image into a canonic form.

3. The method of claim 2, wherein extracting bitmap data from said provided input image comprises:

calculating mapping between the four corners identified from said text region of said input image and four corners of text region of said template;

extracting, from said input image, bitmap data for an outer region corresponding to the outer region of said license plate template image; and

extracting, from said input image, bitmap data for the text region corresponding to the inner region of said license plate template image.

4. The method of claim 3, wherein extracting said bitmap data further comprises:

filtering said outer region bitmap data; and

filtering said text region bitmap data.

5. The method of claim 4, wherein filtering said outer region bitmap data comprises:

utilizing a filter specifically configured to filter said outer region; and

wherein filtering said text region bitmap data comprises utilizing another filter specifically configured to filter said text region.

6. The method of claim 4, the method further comprising:

determining a score for matching said text region bitmap data against said template bitmap text region;

determining another score for matching said outer region bitmap data against said template bitmap outer region; and

determining said license plate input image matches said license plate template image match based on said matching scores.

7. The method of claim 6, the method further comprising:

applying a weighting to said text region matching score; and

applying another weighting to said outer region matching score; and

wherein determining said license plate input image matches said license plate template image comprises determining said license plate input image matches said license plate template image based on said weighted text region matching score and said weighted outer region matching score.

8. The method of claim 1, further comprising:

matching of said plate layout of said license plate image input and said plate layout of said license plate template image; and

determining, based on said matching of said plate layouts, said license plate input image matches said license plate template image match.

9. The method of claim 1, wherein extracting bitmap data from said provided input image comprises:

calculating mapping between text region of said input image and text region of said template;

extracting, from said input image, bitmap data for an outer region corresponding to the outer region of said license plate template image; and

extracting, from said input image, bitmap data for the text region corresponding to the inner region of said license plate template image.

10. A system for verifying automatic license plate recognition results, said system comprising:

a processor;

a data bus coupled to said processor; and

a computer usable medium embodying computer program code, said computer usable medium being coupled to said data bus; and

said computer program code comprising instructions executable by said processor and configured to;

receive an input image of a license plate;

determine a particular license template image associated with said license plate input image by identifying four corners of a text region of said license plate input image, identifying a plate layout based on said identified four corners, and determining said particular license plate template image based on said plate layout;

extract bitmap data from said provided input image;

segment said bitmap data into bitmap data text region and bitmap data outer region;

match said text region bitmap data against a bitmap text region of said license plate image template; and

match said outer region bitmap data against a bitmap outer region of said template.

11. The system of claim 10, wherein said computer program code further comprises instructions executable by said processor and configured to:

normalize said license plate input image and said license plate template image into a canonic form.

12. The system of claim 11, wherein said computer program code further comprises instructions executable by said processor and configured to:

calculate mapping between the four corners identified from said text region of said input image and four corners of text region of said template;

extract, from said input image, bitmap data for an outer region corresponding to the outer region of said license plate template image; and

extract, from said input image, bitmap data for the text region corresponding to the inner region of said license plate template image.

13. The system of claim 12, wherein said computer program code further comprises instructions executable by said processor and configured to:

filter said outer region bitmap data; and

filter said text region bitmap data.

14. The system of claim 13, wherein said computer program code further comprises instructions executable by said processor and configured to:

determine a score for matching said text region bitmap data against said template bitmap text region;

determine a score for matching said outer region bitmap data against said template bitmap outer region; and

determine said license plate input image matches said license plate template image based on said matching scores.

15. The system of claim 14, wherein said computer program code further comprises instructions executable by said processor and configured to:

match said plate layout of said license plate image input and said plate layout of said license plate template image; and

determine, based on said matching of said plate layouts, said license plate input image matches said license plate template image.

16. A computer-usable medium for verifying automatic license plate recognition results, said computer-usable medium embodying computer program code, said computer program code comprising computer executable instructions configured for:

receiving an input image of a license plate to a processing system;

determining a particular license template image associated with said license plate input image by identifying four corners of a text region of said license plate input image, identifying a plate layout based on said identified four corners, and determining said particular license plate template image based on said plate layout;

extracting bitmap data from said provided input image;

segmenting said bitmap data into bitmap data text region and bitmap data outer region;

matching said text region bitmap data against a bitmap text region of said license plate image template; and

matching outer region bitmap data against a bitmap outer region of said template.

17. The computer-usable medium of claim 16, wherein said computer program code comprising computer executable instructions are further configured for:

normalizing said license plate input image and said license plate template image into a canonic form;

calculating mapping between the four corners identified from said text region of said input image and four corners of text region of said template;

extracting, from said input image, bitmap data for an outer region corresponding to the outer region of said license plate template image; and

extracting, from said input image, bitmap data for the text region corresponding to the inner region of said license plate template image.