Segmentation free approach to automatic license plate recognition

• US 9,418,305 B1
• Filed: 04/29/2015
• Issued: 08/16/2016
• Est. Priority Date: 04/29/2015
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Ramesh, Palghat S., Kozitsky, Vladimir, Burry, Aaron Michael
• Primary Examiner(s): Yentrapati, Avinash
• Application Number:
  US14/699,289
• Time in Patent Office: 475 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06V 20/625 License plates
• CPC Class Codes: G06V 20/63 Scene text, e.g. street names
• CPC Class Codes: G06V 30/10 Character recognition
• CPC Class Codes: G06V 30/158 using character size, text spacings or pitch estimation

Abstract

A segmentation free method and system for automatic license plate recognition. An OCR classifier can be swept across an image of a license plate. Characters and their locations can be inferred with respect to the image of the license plate using probabilistic inference based on a Hidden Markov Model (HMM). A language model can be combined with a license plate candidate from the HMM to infer the optimal or best license plate code. The language model can be configured by employing a corpus of license plate codes, wherein the corpus includes a distribution representative of training sets and tests sets.

Claims

1. A method for segmentation free license plate recognition, said method comprising:

sweeping an OCR classifier across an image of a license plate;

inferring characters and their locations with respect to said image of said license plate using probabilistic inference based on at least one Hidden Markov Model (HMM);

combining a language model comprising a Naï

ve Bayes classifier using a MOSCount feature and a codeLength feature, with at least one license plate candidate from said at least one HMM to infer an optimal license plate code; and

identifying a state of origin for said license olate according to said optimal license plate code.

2. The method of claim 1 further comprising configuring said language model by employing a corpus of license plate codes, said corpus having a distribution representative of training sets and tests sets.

3. The method of claim 1 further comprising capturing said image of said license plate with an image-capturing unit.

4. The method of claim 1 further comprising performing a plate localization of said image of said license plate after capturing said image of said license plate.

5. The method of claim 1 further comprising vertically cropping said image of said license plate prior to sweeping said OCR classifier across said image of said license plate.

6. The method of claim 1 wherein inferring characters and their locations with respect to said image of said license plate, further comprises:

performing a character spacing estimation with respect to said characters.

7. The method of claim 6 further comprising vertically cropping said image of said license plate prior to sweeping said OCR classifier across said image of said license plate.

8. The method of claim 6 further comprising performing a plate localization of said image of said license plate after capturing said image of said license plate.

9. A system for segmentation free license plate recognition, said system comprising:

at least one processor; and

a computer-usable medium embodying computer program code, said computer-usable medium capable of communicating with said at least one processor, said computer program code comprising instructions executable by said at least one processor and configured for;

sweeping an OCR classifier across an image of a license plate;

inferring characters and their locations with respect to said image of said license plate using probabilistic inference based on at least one Hidden Markov Model (HMM);

combining a language model comprising a Naï

ve Bayes classifier using a MOSCount feature and a codeLength feature, with at least one license plate candidate from said at least one HMM to infer an optimal license plate code; and

identifying a state of origin for said license olate according to said optimal license plate code.

10. The system of claim 9 wherein said instructions are further provided for configuring said language model by employing a corpus of license plate codes, said corpus having a distribution representative of training sets and tests sets.

11. The system of claim 9 further comprising an image-capturing unit, wherein said image of said license plate is captured with an image-capturing unit and wherein said image-capturing unit communicates with said at least one processor.

12. The system of claim 11 wherein said instructions are further configured for performing a plate localization of said image of said license plate after capturing said image of said license plate via said image-capturing unit.

13. The system of claim 9 wherein said instructions are further configured for vertically cropping said image of said license plate prior to sweeping said OCR classifier across said image of said license plate.

14. The system of claim 9 wherein said instructions for inferring characters and their locations with respect to said image of said license plate, further comprise instructions configured for performing a character spacing estimation with respect to said characters.

15. The system of claim 11 wherein said instructions are further configured for vertically cropping said image of said license plate prior to sweeping said OCR classifier across said image of said license plate.

16. The system of claim 15 wherein said instructions for inferring characters and their locations with respect to said image of said license plate, further comprise instructions configured for performing a character spacing estimation with respect to said characters.

17. A system for segmentation free license plate recognition, said system comprising:

an image-capturing unit that captures an image of a license plate;

an OCR classifier, wherein said OCR classifier is swept across said image of said license plate;

at least one Hidden Markov Model (HMM) wherein characters and their locations with respect to said image of said license plate are inferred using probabilistic inference based on said at least one HMM;

a language model comprising a Naï

ve Bayes classifier using a MOSCount feature and a codeLength feature, that is combined with at least one license plate candidate from said at least one HMM to infer an optimal license plate code; and

a state identification module that identifies a state of origin for said license olate according to said optimal license plate code.

18. The system of claim 17 wherein said language mode is configured by employing a corpus of license plate codes, said corpus having a distribution representative of training sets and tests sets.

19. The system of claim 17 wherein a plate localization of said image of said license plate is performed after capturing said image of said license plate.

20. The system of claim 17 wherein said image of said license plate prior is vertically cropped prior to sweeping said OCR classifier across said image of said license plate.