VEHICLE LICENSE PLATE RECOGNITION METHOD AND SYSTEM THEREOF

• US 20120106802A1
• Filed: 10/26/2011
• Published: 05/03/2012
• Est. Priority Date: 10/29/2010
• Status: Active Grant

• Current Assignee: National Chiao Tung University (Government of The Republic of China)
• Original Assignee: National Chiao Tung University (Government of The Republic of China)
• Inventors: HSIEH, JUN-WEI, CHEN, SIN YU
• Granted Patent: US 8,509,486 B2
• Time in Patent Office: Days
• US Class Current: 382/105
• CPC Class Codes: G06V 20/625 License plates
• CPC Class Codes: G06V 20/63 Scene text, e.g. street names

Abstract

A vehicle license plate recognition method and a system thereof are disclosed. A region where a vehicle license plate image exists is detected according to the edge densities of an input image and a vehicle license plate specification. A text area of the vehicle license plate image is divided into a plurality of character images. The character images are binarized to obtain a plurality of binarized character images. A plurality of characters is recognized from the binarized character images. The characters are recombined to form a character string. The abovementioned steps are repeated to obtain a new character string from another image of the same vehicle, which is captured at a next time point. The character string is compared with the new character string character by character to obtain a comparison result for verifying reliability of recognition through a voting technique.

Claims

1. A vehicle license plate recognition method comprising:

Step (a);

converting an image of a vehicle into a gray-level image, calculating edge densities of said gray-level image, and defining a region where a vehicle license plate image exists according to said edge densities and a vehicle license plate specification;

Step (b);

detecting a text area of said vehicle license plate image, partitioning said text area into a plurality of character images, and binarizing said character images to obtain a plurality of binarized character images;

Step (c);

recognizing a plurality of characters from said binarized character images;

Step (d);

recombining said characters into a character string of a vehicle license plate and outputting said character string; and

Step (e);

receiving a next image of said vehicle, which is captured at a different time point, and repeating steps of from said Step (a) to said Step (d) to obtain a new character string from said next image;

comparing said character string with said new character string character by character to obtain a comparison result for verifying reliability of recognition.

2. The vehicle license plate recognition method according to claim 1, wherein said image is captured by an image capture device.

3. The vehicle license plate recognition method according to claim 1, wherein said gray-level is divided into a plurality of sections, and wherein an integral image technology is used to calculate said edge densities of each of said sections.

4. The vehicle license plate recognition method according to claim 3, wherein said integral image technology accumulates the gradient magnitudes of each pixel of said gray-level image, and wherein said edge density equals to a sum of said gradient magnitudes divided by a region size.

5. The vehicle license plate recognition method according to claim 1, wherein regions having a higher value of said edge density are appointed as candidate license plate areas, and wherein a predetermined threshold edge density is used to exclude said candidate license plate areas having said edge density lower than said predetermined threshold edge density.

6. The vehicle license plate recognition method according to claim 5, wherein said candidate license plate areas having said edge density higher than said predetermined threshold edge density are merged according to a merging condition to form a plurality of merged candidate license plate areas, and wherein said merged candidate license plate areas that do not meeting said vehicle license plate specification are excluded.

7. The vehicle license plate recognition method according to claim 6, wherein said merging condition is that each two said candidate license plate areas that overlap are merged to form one said merged candidate license plate area.

8. The vehicle license plate recognition method according to claim 1, wherein said vehicle license plate specification is a predetermined width-to-width ratio.

9. The vehicle license plate recognition method according to claim 1, wherein a vertical projection technology is used to obtain positive projection values and negative projection values of said vehicle license plate image, and wherein maximums of said positive projection values and said negative projection values are used to define a top boundary and a bottom boundary of said text area, and wherein a horizontal gray-level projection is performed on a region between said top boundary and said bottom boundary to find out a plurality of peaks, and wherein said peaks are used to define a left boundary and a right boundary of each character image in said text area, and wherein said character image is partitioned according to said top boundary, said bottom boundary, said left boundary and said right boundary.

10. The vehicle license plate recognition method according to claim 9, wherein said vehicle license plate image is divided into an upper half and a lower half, and wherein said vertical projection technology calculates maximums of said positive projection values and said negative projection values of said upper half and said lower half to determine a top boundary and a bottom boundary of said text area in a cases of white characters in a black background or in a case of black characters in a white background and determine height of said characters, and wherein positions of said peaks are used to determine whether said text area is in a case of white characters in a black background or in a case of black characters in a white background.

11. A vehicle license plate recognition system comprising

a license plate detection module detecting an image, converting said image into a gray-level image, calculating edge densities of said gray-level image, and defining a region where a vehicle license plate image exists according to said edge densities and a vehicle license plate specification;

a character partition module connected with said license plate detection module, partitioning said vehicle license plate image into a plurality of character images, and binarizing said character images to obtain a plurality of binarized character images;

a character recognition module connected with said character partition module, recognizing characters from said binarized character images;

a character recombination module connected with said character recognition module, recombining said characters into a character string of a vehicle license plate, and outputting said character string; and

a voting module connected with said character recognition module, obtaining a new character string generated in a next time point, comparing said character string with said new character string character by character to generate a comparison result for verifying reliability of recognition.

12. The vehicle license plate recognition system according to claim 11, wherein an image capture device captures said image and sends said image to said license plate detection module.

13. The vehicle license plate recognition system according to claim 11, wherein said license plate detection module divides said gray-level image into a plurality of sections, and an integral image technology is used to calculate edge densities of all said sections.

14. The vehicle license plate recognition system according to claim 13, wherein said integral image technology accumulates the gradient magnitudes of pixels of said gray-level image, and wherein said edge density equals to a sum of said gradient magnitudes divided by a region size.

15. The vehicle license plate recognition system according to claim 11, wherein said merging condition is that each two said candidate license plate areas that overlap are merged to form one said merged candidate license plate area.

16. The vehicle license plate recognition system according to claim 11, wherein said license plate detection module uses a vertical projection technology to obtain positive projection values and negative projection values of said vehicle license plate image, uses maximums of said positive projection values and said negative projection values to define a top boundary and a bottom boundary of a text area of said vehicle license plate image, performs a horizontal gray-level projection on an area between said top boundary and said bottom boundary to find out a plurality of peaks, and defines a left boundary and a right boundary of each said character in said text area according to said peaks, and wherein said character partition module partitions said vehicle license plate image into said character images according to said top boundary, said bottom boundary, said left boundary and said right boundary.

17. The vehicle license plate recognition system according to claim 16, wherein said vehicle license plate image is divided into an upper half and a lower half, and wherein said vertical projection technology calculates maximums of said positive projection values and said negative projection values of said upper half and said lower half to determine a top boundary and a bottom boundary of said text area in a cases of white characters in a black background or in a case of black characters in a white background and determine height of said characters, and wherein positions of said peaks are used to determine whether said text area is in a case of white characters in a black background or in a case of black characters in a white background.

18. The vehicle license plate recognition system according to claim 11, wherein said character partition module performs binarization on a block between two neighboring peaks of each said character image to obtain said binarized character images corresponding to said character images.