Method and system of identifying one or more features represented in a plurality of sensor acquired data sets

• US 8,457,408 B2
• Filed: 08/31/2007
• Issued: 06/04/2013
• Est. Priority Date: 09/01/2006
• Status: Active Grant

• Current Assignee: SenSen Networks Group Pty., Ltd.
• Original Assignee: SenSen Networks Pty., Ltd. (SenSen Networks Ltd.)
• Inventors: Challa, Subhash
• Primary Examiner(s): Mehta, Bhavesh
• Assistant Examiner(s): Park, Edward
• Application Number:
  US12/439,531
• Publication Number:
  US 20100054539A1
• Time in Patent Office: 2,104 Days
• Field of Search: 382/105, 382/190, 382/208, 382/181, 382/155, 382/161, 382/100, 382/104
• US Class Current: 382/190
• CPC Class Codes: G06F 18/256 of results relating to different input data, e.g. multimodal recognition
• CPC Class Codes: G06V 10/811 the classifiers operating on different input data, e.g. multi-modal recognition
• CPC Class Codes: G06V 20/625 License plates
• CPC Class Codes: G06V 20/63 Scene text, e.g. street names

Abstract

A method and system for identifying one or more features represented in a plurality of sensor acquired data sets is described. The method and apparatus is particularly useful in automatic license plate recognition applications, where the sensor acquired data sets are data obtained from one or more digital cameras. This is achieved by determining a first probability of the identity of the one or more features (eg alphanumeric characters) from a first one of the data sets; determining a second probability of the identity of the one or more features from a second one of the data sets; and, using data fusion techniques, fusing the determined first and second probabilities to provide a fused probability. This fused probability is used to identify the one or more features from data sets.

Claims

1. A method of identifying one or more features represented in a plurality of sensor acquired data sets, the method comprising:

determining a first probability of the identity of the one or more features from a first one of the data sets;

determining a second probability of the identity of the one or more features from a second one of the data sets;

fusing the determined first and second probabilities to provide a fused probability;

after the fusing;

determining p(n);

fusing p(fused_n-2) with p(n) to provide p(fused_n-1),

where p(n) is a probability of the one or more features being present in an n^th one of the images, and

p(fused_n-2) and p(fused_n-1) are (n−

2)^th and (n−

1)^th fused probabilities, respectively; and

repeating the determining of p(n) and fusing p(fused_n-2) with p(n) for m times, where m=n−

2, and n is an integer greater than 2; and

identifying the one or more features from the fused probability, wherein the identifying uses (fused_n-1).

2. The method of claim 1, wherein an algorithm selected from the group comprising Bayesian Fusion, Distributed Data Fusion, Dempster-Shafer Fusion, Fuzzy Fusion, Random Sets Based Fusion, Voting and Dezert Samaranche Fusion is used in the fusing to fuse the determined probabilities.

3. The method of claim 1 further comprising, after the fusing:

determining a third probability of the identity of the one or more features being present in a third one of the data sets; and

fusing the first mentioned fused probability and the third probability to provide a second fused probability,

wherein the identifying uses the second fused probability.

4. The method of claim 1, wherein each of the probabilities are probability distributions representing probabilities of the presence of each one of a set of predetermined said features in their respective data sets and wherein the identifying is performed by determining which of the predetermined features has the highest probability from the probability distribution of the fused probability.

5. The method of claim 1, wherein the one or more features comprise alphanumeric characters, and wherein each of the plurality of data sets comprise different respective representations of a vehicle license plate, the plate displaying one or more of the alphanumeric characters.

6. The method of claim 5, wherein when the plate displays a combination of at least two alphanumeric characters, the method is repeated to identify a first one of the two characters and then a second one of the two characters from each of the at least two data sets.

7. The method of claim 5 wherein, prior to determining the first probability, the number of alphanumeric characters present on the plate is determined, and the remainder of the method is performed for each alphanumeric character determined as present on the plate.

8. The method of claim 1 wherein the plurality of data sets represent respective images comprising a representation of a license plate and the one or more features comprise respective location(s) of the license plate in respective said images.

9. The method of claim 8 wherein the first and subsequent probability(ies) are first and subsequent probability density functions representing the probability of the license plate being in a location in the image.

10. The method of claim 8 wherein, prior to determining the first probability, the images used in the method are subjected to black and white image thresholding and the probability determinations are performed on the thresholded black and white images.

11. The method of claim 8 wherein the first and subsequent probabilities are determined using one or more parameters of the following group:

size of a potential representation of the license plate compared to other potential representation(s) of the license plate;

shape of the potential representation of the license plate;

similarities of size and/or shape of the potential representation of the license plate in one of the images compared to the size and/or shape of another potential representation in another of the images; and

the position of the potential representation of the license plate in one of the images relative to the position of another potential representation in another of the images.

12. The method of claim 1 implemented by a computer.

13. A computer program product stored on a non-transitory computer-readable medium and configured to cause a computer to perform the method of claim 1.

14. A non-transitory computer readable medium providing the computer program of claim 13.

15. A system for identifying one or more features represented in a plurality of sensor acquired data sets, the system comprising:

one or more sensors configured to acquire a plurality of data sets;

a device in communication with the one or more sensors configured to acquire the data sets from the one or more sensors;

a calculating device configured to perform the method of claim 1.

16. A method of identifying a vehicle registration number from a plurality of images comprising representations of a vehicle registration plate, the method comprising:

determining a first probability distribution of alphanumeric characters present in a portion of one of the images;

determining a second probability distribution of alphanumeric characters present in a portion of another of the images, the portion of the other of the images corresponding to the portion of the one of the images;

fusing the first and second probability distributions to provide a fused probability distribution of alphanumeric characters in relation to the one and the other images;

after the fusing;

determining p(n);

fusing p(fused_n-2) with p(n) to provide p(fused_n-1), where

p(n) is a probability density function in relation to alphanumeric character being present in an n^th one of the images, and

p(fused_n-2) and p(fused_n-1) are (n−

2)^th and (n−

1)^th fused probability density functions, respectively; and

repeating the determining of p(n) and fusing p(fused_n-2) with p(n) for m times, where m=n−

2, and n is an integer greater than 2; and

identifying which alphanumeric character is present in the respective portions of the one and the other images by identifying from the fused probability distribution the alphanumeric character having the highest probability of being present, wherein the identifying comprises using (fused_n-1) to identify the alphanumeric character with the highest probability of being present.

17. The method of claim 16, wherein an algorithm selected from the group of algorithms comprising Bayesian Fusion, Distributed Data Fusion, Dempster-Shafer Fusion, Fuzzy Fusion, Random Sets Based Fusion, Voting and Dezert Samaranche Fusion is used in the fusing to fuse the first and second probability distributions.

18. The method of claim 16 further comprising, prior to determining the first probability distribution, determining the number of alphanumeric characters present on the vehicle registration plate, and performing the method for each alphanumeric character present on the vehicle registration plate.

19. A system for identifying one or more features represented in a plurality of sensor acquired data sets, the system comprising:

one or more sensors configured to acquire a plurality of data sets;

a device in communication with the one or more sensors configured to acquire the data sets from the one or more sensors;

a calculating device configured to;

determine a probability of the identity of one of the features from at least two said data sets;

fuse the determined probabilities to provide a fused probability;

after the fusing;

fuse p(fused_n-1) with p(n) to obtain p(fused_n-1), where p(n) is a probability density function in relation to alphanumeric characters being present in an n^th one of the images, and p(fused_n-2) and p(fused_n-1) are (n−

2)^th and (n−

1)^th fused probability density functions, respectively; and

repeat the determining of p(n) and the fusing p(fused_n-2) with p(n) for m times, where m=n−

2, and n is an integer greater than 2; and

an identifying device configured to identify the one or more features from the fused probability, using (fused_n-1) to identify the alphanumeric character with the highest probability of being present.

20. The system of claim 19 wherein an algorithm selected from the group of algorithms comprising Bayesian Fusion, Distributed Data Fusion, Dempster-Shafer Fusion, Fuzzy Fusion, Random Sets Based Fusion, Voting and Dezert Samaranche Fusion is used to fuse the determined probabilities.