Number plate recognition system

• US 5,315,664 A
• Filed: 05/26/1993
• Issued: 05/24/1994
• Est. Priority Date: 12/02/1989
• Status: Expired due to Fees

• Current Assignee: YOZAN, Inc.
• Original Assignee: Ezel, Inc.
• Inventors: Kumagai, Ryohei
• Primary Examiner(s): Mancuso, Joseph
• Application Number:
  US08/067,166
• Time in Patent Office: 363 Days
• Field of Search: 382/1, 382/30, 382/9, 382/48, 382/50, 382/54, 382/25, 382/51, 382/52, 382/53, 382/27
• US Class Current: 382/105
• CPC Class Codes: G06V 10/40 Extraction of image or video features
• CPC Class Codes: G06V 30/1478 of characters or characters lines

Abstract

A number plate recognition system has structure for inputting an image; structure for calculating outlines of each configuration included in an input image, as well as for recognizing a configuration with the maximal outline size as the plate area. The areas in an image except the above plate area are masked, and the numbers are determined by using Euler numbers after calculating Euler numbers for each configuration within the above plate area.

Claims

1. A number plate recognition system comprising:

a video camera which obtains an image which includes a number plate to be recognized;

processing means, receiving said image, for;

1) determining an outline of the number plate in both the vertical and the horizontal directions;

2) labeling portions of the image to determine continuous configurations within the image, each continuous configuration within the image being classified as a group having a different labelling number,

3) determining sizes of the image portions bounded by the labeling to determine a configuration,

4) setting one of the configurations having a maximal diameter which is a largest size diameter, based on said sizes, as a diameter of the number plate,

5) determining variances for each said pixel within said groups;

6) dividing the pixels into subgroups according to their densities;

7) calculating a threshold whereby a ratio between variance within a group and variance between groups is maximum; and

8) using said threshold to binarize said pixels; and

means for calculating values within an area defined by said maximal diameter from which numbers on said number plate can be calculated.

2. A system as in claim 1, wherein said grouping is made by densities.

3. A number plate recognition system as in claim 1 wherein said processing means is also for calculating differentials for said plurality of pixels in said vertical and horizontal directions and using values of said differentials to determine an outline of the number plate in both the vertical and the horizontal directions.

4. A method of recognizing a number plate recognition system comprising;

means for obtaining an image which includes a number plate to be recognized;

processing means receiving said image, for;

1) determining densities of portions of the image and labelling said densities, to determine continuous configurations within the image, each continuous configuration within the image being classified as a group having a different labelling number,

2) determining sizes of the image portions bounded by the labeling to determine a configuration,

3) setting one of the configurations having a maximal diameter which is a largest diameter found, based on said sizes, as a diameter of the number plate;

4) determining variances for each pixel within said groups;

6) dividing the pixels into two subgroups according to their densities;

7) calculating a threshold whereby a ratio between variance within a group and variance between groups is maximum; and

8) using said threshold to binarize said pixels; and

means for calculating values within an area defined by said maximal diameter from which numbers on said number plate can be calculated.

5. A number plate recognition system as in claim 4 wherein said processing means is also for calculating differentials for said plurality of pixels in said vertical and horizontal directions and using values of said differentials to determine an outline of the number plate in both the vertical and the horizontal directions.