METHODS AND SYSTEMS FOR CHARACTER SEGMENTATION IN AUTOMATED LICENSE PLATE RECOGNITION APPLICATIONS

• US 20140307924A1
• Filed: 04/15/2013
• Published: 10/16/2014
• Est. Priority Date: 04/15/2013
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Fillion, Claude S., Burry, Aaron Michael, Kozitsky, Vladimir
• Granted Patent: US 9,122,953 B2
• Time in Patent Office: Days
• US Class Current: 382/105
• CPC Class Codes: G06V 10/26 Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
• CPC Class Codes: G06V 20/625 License plates

Abstract

Methods and systems for character segmentation in an automatic license plate recognition application. One or more images of a license plate are acquired. Then, a pixel-level importance may be calculated with respect to the image(s) of the license plate based on information within the image, such as gradient information and raw grayscale information. A seam selection can be then applied with respect to the pixel-level importance map and the image(s) by enforcing constraints based on known characteristics of license plates in order to provide for character segmentation with respect to the image(s) of the license plate.

Claims

1. A method for character segmentation with respect to an image of a license plate, said method comprising:

acquiring at least one image of a license plate;

calculating a pixel-level importance map with respect to said at least one image of said license plate and based on images attributes associated with said at least one image; and

applying a seam selection with respect to said pixel-level importance map and said at least one image by enforcing at least one constraint based on known characteristics of license plates in order to provide for a character segmentation with respect to said at least one image of said license plate.

2. The method of claim 1 wherein said image attributes comprise at least one of gradient information and raw grayscale information.

3. The method of claim 1 wherein said at least one constraint comprises at least one of a spacing constraint and a seam grouping.

4. The method of claim 2 further comprising highlighting the importance of characters within said at least one image to support said character segmentation individually or with a hybrid map with respect to said at least one image.

5. The method of claim 4 wherein said hybrid map comprises an importance map M wherein a weighted combination of gradient magnitude G with respect to said gradient information and a complement of image intensity I_c with respect to said raw grayscale information are indicated by the following equation:

M(i,h)=α

G(i,j)+β

I_c(i,j) where α

=5 and β

=1.

6. The method of claim 1 further comprising compiling said at least one constraint based on said known characteristics of license plates.

7. The method of claim 2 further comprising combining said gradient information with said raw grayscale information to target said character segmentation to between letters identified in said at least one image of said license plate rather than within a letter thereof to minimize false seams with respect to a seam selection.

8. The method of claim 7 further comprising adding a spacing constraint to said at least one constraint to assist in minimizing said false seams with respect to said seam selection.

9. The method of claim 1 wherein:

said image attributes comprise at least one of gradient information and raw grayscale information; and

said at least one constraint comprises at least one of a spacing constraint and a seam grouping.

10. A system for character segmentation with respect to an image of a license plate, 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, said computer program code comprising instructions executable by said processor and configured for;

acquiring at least one image of a license plate;

calculating a pixel-level importance map with respect to said at least one image of said license plate and based on images attributes associated with said at least one image; and

applying a seam selection with respect to said pixel-level importance map and said at least one image by enforcing at least one constraint based on known characteristics of license plates in order to provide for a character segmentation with respect to said at least one image of said license plate.

11. The system of claim 10 wherein said image attributes comprise at least one of gradient information and raw grayscale information.

12. The system of claim 10 wherein said at least one constraint comprises at least one of a spacing constraint and a seam grouping.

13. The system of claim 11 wherein said instructions are further configured for highlighting the importance of characters within said at least one image to support said character segmentation individually or with a hybrid map with respect to said at least one image.

14. The system of claim 13 wherein said hybrid map comprises an importance map M wherein a weighted combination of gradient magnitude G with respect to said gradient information and a complement of image intensity I_c with respect to said raw grayscale information are indicated by the following equation:

M(i,j)=α

G(i,j)+β

I_c(i,j) where α

=5 and β

=1.

15. The system of claim 10 wherein said instructions are further configured for compiling said at least one constraint based on said known characteristics of license plates.

16. The system of claim 11 wherein said instructions are further configured for combining said gradient information with said raw grayscale information to target said character segmentation to between letters identified in said at least one image of said license plate rather than within a letter thereof to minimize false seams with respect to a seam selection.

17. The system of claim 16 wherein said instructions are further configured for adding a spacing constraint to said at least one constraint to assist in minimizing said false seams with respect to said seam selection.

18. A processor-readable medium storing computer code representing instructions to cause a process for character segmentation with respect to an image of a license plate, said computer code comprising code to:

acquire at least one image of a license plate;

calculate a pixel-level importance map with respect to said at least one image of said license plate and based on images attributes associated with said at least one image; and

apply a seam selection with respect to said pixel-level importance map and said at least one image by enforcing at least one constraint based on known characteristics of license plates in order to provide for a character segmentation with respect to said at least one image of said license plate.

19. The processor-readable medium of claim 18 wherein said image attributes comprise at least one of gradient information and raw grayscale information.

20. The processor-readable medium of claim 18 wherein said at least one constraint comprises at least one of a spacing constraint and a seam grouping.