Method for automatic license plate recognition using adaptive feature set

1. A method for determining a confidence level to be used in identifying a vehicle comprising:

receiving, by a processing device, at least one vehicle image;

determining, by the processing device, a license plate number, wherein the license plate number comprises a plurality of characters;

determining, by the processing device, a confidence score for each of the plurality of characters, thereby resulting in a plurality of confidence scores for the vehicle image;

determining, by the processing device, an associated confidence level based upon the at least one vehicle image and the confidence score for each of the plurality of characters;

extracting, by the processing device, auxiliary data from the at least one vehicle image, wherein the auxiliary data comprises information related to physical details of the vehicle;

corresponding, by the processing device, the extracted auxiliary data and a set of stored auxiliary data; and

updating, by the processing device, the associated confidence level by updating the confidence score for each of the plurality of characters to produce an updated confidence level based upon the correspondence of the extracted auxiliary data and the set of stored auxiliary data,

wherein the extracted auxiliary data and the plurality of confidence scores each comprise an associated weight, and wherein the extracted auxiliary data comprises a higher associated weight than the plurality of confidence scores if the vehicle image comprises a high likelihood of being inaccurate due to various conditions during image acquisition.

2. The method of claim 1 , wherein the determining a license plate number and associated confidence level further comprises performing an optical character recognition (OCR) identification algorithm on the at least one vehicle image.

3. The method of claim 2 , wherein performing the OCR identification algorithm comprises:

extracting a license plate image from the vehicle image, wherein the license plate image comprises the plurality of characters;

segmenting each of the plurality of characters in the license plate image;

comparing each of the segmented characters against a set of standard character images; and

assigning a confidence score to each of the segmented characters.

4. The method of claim 1 , wherein the auxiliary data comprises at least one of vehicle color, vehicle shape, vehicle design elements near the license plate, distinguishing marks, physical attributes, customized features, vehicle make, vehicle model, and model badges and emblems.

5. The method of claim 1 , wherein the auxiliary data comprises at least one of time of day, day of the week, and location.

6. The method of claim 1 , further comprising updating the set of stored auxiliary data upon updating the confidence level.

7. The method of claim 1 , further comprising comparing the updated confidence level against a confidence threshold.

8. The method of claim 7 , further comprising:

reporting the license plate number if the updated confidence level is higher than the confidence threshold; and

obtaining human interpretation if the updated confidence level is lower than the confidence threshold.

9. The method of claim 1 , wherein the set of stored auxiliary data is stored on a non-transitory processor readable storage medium operably connected to the processing device.

10. A method for determining a confidence level to be used in identifying a vehicle comprising:

receiving, by a processing device, at least one vehicle image;

extracting, by the processing device, a license plate image from the at least one vehicle image;

determining, by the processing device, a license plate number, wherein the license plate number comprises a plurality of characters;

determining, by the processing device, a confidence score for each of the plurality of characters, thereby resulting in a plurality of confidence scores for the vehicle image;

determining, by the processing device, an associated confidence level based upon the license plate image and the confidence score for each of the plurality of characters;

extracting, by the processing device, auxiliary data from the at least one vehicle image, wherein the auxiliary data comprises information related to physical details of the vehicle;

corresponding, by the processing device, the extracted auxiliary data and a set of stored auxiliary data;

updating, by the processing device, the associated confidence level by updating the confidence score for each of the plurality of characters to produce an updated confidence level based upon the correspondence of the extracted auxiliary data and the set of stored auxiliary data; and

comparing, by the processing device, the updated confidence level against a confidence threshold,

wherein the extracted auxiliary data and the plurality of confidence scores each comprise an associated weight, and wherein the extracted auxiliary data comprises a higher associated weight than the plurality of confidence scores if the vehicle image comprises a high likelihood of being inaccurate due to various conditions during image acquisition.

11. The method of claim 10 , wherein the determining a license plate number and associated confidence level further comprises performing an optical character recognition (OCR) identification algorithm on the license plate image.

12. The method of claim 11 , wherein performing the OCR identification algorithm comprises:

identifying the plurality of characters in the license plate image;

segmenting each of the plurality of characters in the license plate image;

comparing each of the segmented characters against a set of standard character images; and

assigning a confidence score to each of the segmented characters.

13. The method of claim 10 , wherein the auxiliary data comprises at least one of vehicle color, vehicle shape, vehicle design elements near the license plate, distinguishing marks, physical attributes, customized features, vehicle make, vehicle model, and model badges and emblems.

14. The method of claim 10 , wherein the auxiliary data comprises at least one of time of day, day of the week, and location.

15. The method of claim 10 , further comprising:

reporting the license plate number if the updated confidence level is higher than the confidence threshold; and

obtaining human interpretation if the updated confidence level is lower than the confidence threshold.

16. A method for determining a confidence level to be used in identifying a vehicle comprising:

receiving, by a processing device, a vehicle image;

extracting, by the processing device, a license plate image from the at least one vehicle image;

determining, by the processing device, a license plate number, wherein the license plate number comprises a plurality of characters;

determining, by the processing device, a confidence score for each of the plurality of characters, thereby resulting in a plurality of confidence scores for the vehicle image;

determining, by the processing device, an associated confidence level based upon the license plate image and the confidence score for each of the plurality of characters;

comparing the associated confidence level against a confidence threshold; and

if the associated confidence level is below the confidence threshold:

extracting, by the processing device, auxiliary data from the at least one vehicle image, wherein the auxiliary data comprises information related to physical details of the vehicle,

corresponding, by the processing device, the extracted auxiliary data and a set of stored auxiliary data, and

updating, by the processing device, the associated confidence level by updating the confidence score for each of the plurality of characters to produce an updated confidence level based upon the correspondence of the extracted auxiliary data and the set of stored auxiliary data,

wherein the extracted auxiliary data and the plurality of confidence scores each comprise an associated weight, and wherein the extracted auxiliary data comprises a higher associated weight than the plurality of confidence scores if the vehicle image comprises a high likelihood of being inaccurate due to various conditions during image acquisition.

17. The method of claim 16 , wherein the determining a license plate number and associated confidence level further comprises performing an object character recognition (OCR) identification algorithm on the license plate image.

18. The method of claim 17 , wherein performing the OCR identification algorithm comprises:

identifying the plurality of characters in the license plate image;

segmenting each of the plurality of characters in the license plate image;

comparing each of the segmented characters against a set of standard character images; and

assigning a confidence score to each of the segmented characters.

19. The method of claim 16 , further comprising:

reporting the license plate number if the updated confidence level is higher than the confidence threshold; and

obtaining human interpretation if the updated confidence level is lower than the confidence threshold.