Multi-query privacy-preserving parking management system and method

• US 9,792,301 B2
• Filed: 09/26/2014
• Issued: 10/17/2017
• Est. Priority Date: 09/26/2014
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Inventors: Rodriguez-Serrano, Jose Antonio, Perronnin, Florent
• Primary Examiner(s): Trujillo, James
• Assistant Examiner(s): TESSEMA, AIDA Z
• Application Number:
  US14/497,417
• Publication Number:
  US 20160092473A1
• Time in Patent Office: 1,117 Days
• Field of Search: None
• CPC Class Codes: G06F 16/5846 using extracted text
• CPC Class Codes: G06Q 20/145 Payments according to the detected use or quantity
• CPC Class Codes: G07B 15/02 taking into account a variable factor such as distance or time, e.g. for passenger transport, parking systems or car rental systems G07B15/06 takes precedence; taximeters G07B13/00; parking meters per se G07F17/24; car rental systems per se G07B15/00
• CPC Class Codes: G07F 17/24 for parking meters

Abstract

A privacy-preserving multi-query parking management system and method. An image of a vehicle with respect to a specific part (e.g., license plate) can be captured when the vehicle enters via an entry point. An image signature can be computed by embedding the image into a descriptor space and the image signature can be stored in a database together with relevant metadata. An image of a vehicle with respect to the specific part can be captured when the vehicle exits via at least one exit point and an image signature can be computed. An entry-exit association with respect to the vehicle can be performed by searching for a most similar image signature in the database. Additionally, given a textual query (e.g., a license plate number), a signature can be computed by embedding the text into the same descriptor space.

Claims

1. A privacy-preserving multi-query parking management system, comprising:

at least one image capturing unit among a plurality of image capturing units;

at least one entry point wherein at least one license plate image signature is captured with said at least one image capturing unit, computed, and stored in a database of image signatures, said image signatures including said at least one license plate image signature, said at least one image capturing unit having built-in integrated functions including image processing, data formatting, data compression, imager-positioning, range-finding, and a flash bulb, wherein said at least one image capturing unit is associated with said at least one entry point;

at least one exit point wherein at least one license plate image signature is captured via at least one other image capturing unit among said plurality of image capturing units, computed, and matched to said database, wherein said at least one license plate image signature is computed by embedding an image associated with said at least one license plate image signature into a descriptor space, and wherein said at least one license plate image signature is stored in said database together with metadata relevant to said at least one license plate image signature, wherein said at least one other image capturing unit is associated with said at least one exit point;

at least one electronic kiosk having a touch screen display that enables a query-by-license plate number with respect to said image signatures stored in said database based on a license plate number and license plate image association, said electronic kiosk including payment processing with respect to a vehicle parking;

wherein said descriptor space includes at least one patch descriptor, and wherein said at least one license plate image signature is computed and stored in said database as at least one Fisher Vector image signature;

wherein a statistic with respect to said at least one patch descriptor is computed and aggregated utilizing a Fisher Vector Network; and

wherein said at least one license plate image signature is rendered more descriptive by applying a projection obtained utilizing a low-rank Mahalanobis metric.

2. The system of claim 1 wherein said at least one license plate signature is queried based on an image-to-image query.

3. The system of claim 1 wherein said a least one license plate signature is queried based on a text-to-image query.

4. The system of claim 1 wherein said license plate signature is queried based on an image-to-text query.

5. The system of claim 1 wherein said kiosk comprises an electronic parking kiosk.

6. The system of claim 1 wherein said at least one license plate image signature is computed and stored in said database as at least one spatial pyramid of characters text signature.

7. The system of claim 1 wherein said database is queried to determine a lost ticket.

8. The system of claim 1 wherein said database is queried to determine a presence of a vehicle with respect to a parking location associated with said at least one image capturing unit or said at least one other capturing unit among said plurality of image capturing units.

9. The system of claim 8 wherein said database further comprises a subscriber database to which subscribers are added to allow free entry and/or free exist respectively through said at least one entry point and/or said at least one exit point and wherein said system performs an image-to-image query, a text-to-image query, and an image-to-text query.

10. A mufti-query privacy-preserving parking management system, comprising:

at least one image capturing unit among a plurality of image capturing units, said at least one image capturing unit having built-in integrated functions including image processing, data formatting, data compression, imager-positioning, range-finding, and a flash bulb;

a database that stores information collected at entry and enables query-by-image and query-by-license plate number functionalities;

at least one entry point wherein when a vehicle enters, at least one image of said vehicle is captured and at least one image signature associated with at least one image of said vehicle computed by embedding said image into a descriptor space and stored in said database, said descriptor space including at least one patch descriptor and wherein said at least one license plate image signature is computed and stored in said database as at least one Fisher Vector image signature, said descriptor space comprising a same descriptor space as a descriptor space used for an image associated with said at least one image signature;

at least one exit point wherein when said vehicle proceeds to exit through said at least one exit point, at least one image of said vehicle is captured and an entry-exit association performed with respect to said vehicle by searching for a signature most similar to said at least one image signature in said database; and

at least one kiosk that enables said query-by-license-plate number functionalities with respect to said at least one image signature database and wherein said at least one kiosk further enables payment operations including marking a parking ticket as paid and including a payment timestamp, wherein said payment timestamp is stored in said database.

11. The system of claim 10 wherein said at least one image signature is stored in said database with relevant metadata wherein said relevant data includes an entry time associated with said vehicle entering said at least on entry point and a vehicle type of said vehicle.

12. The system of claim 11 wherein said system performs an image-to-image query, a text-to-image query, and an image-to-text query.

13. The system of claim 12 wherein said at least one image signature is computed and stored in, said database as at least one spatial pyramid of characters text signature.

14. A multi-query parking management method, said method comprising:

computing a first image signature with respect to an image of a vehicle part captured by said at least one image capturing unit among a plurality of image capturing units at an entry point by embedding said image into a descriptor space and storing said first image signature in a database together with relevant metadata, said at least one image capturing unit having built-in integrated functions including image processing, data formatting, data compression, imager-positioning, range-finding, and a flash bulb;

calculating a second image signature with respect to an image of said vehicle part captured at an exit point by at least one image capturing unit having built-in integrated functions including image processing, data formatting, data compression, imager-positioning, range-finding, and a flash bulb and storing said second image signature in said database;

performing an entry-exit association by searching and retrieving a most similar image signature with respect to said second image signature from said database;

computing a text signature from a textual query rendered at said electronic kiosk and facilitated through a touch screen display associated with said electronic kiosk;

utilizing said text signature to assist in confirming said vehicle; and

querying said database via an electronic kiosk that communicates with said database to obtain information about a vehicle associated with said vehicle part, and wherein computing said text signature further comprises synthesizing said license plate image for a corresponding license plate number wherein embedding of said license plate number comprises an image signature of said synthesized plate and wherein said image of said vehicle part comprises a license plate image.

15. The method of claim 14 wherein computing said text signature further comprises computing a spatial pyramid of characters and developing a mapping between a space of said text signature and a space of an image signature utilizing a projection matrix.

16. The method of claim 15 further comprising performing a matching between said image signatures and a matching between said image signatures and said text signature by measuring a similarity and a distance between two vectorial representations.

17. The method of claim 16 wherein computing said first image signature or said second image signature, further comprises:

extracting a set of local patches from said vehicle part image and describing said patches with said descriptor;

computing and aggregating a set of statistics with respect to each patch descriptor using a Fisher Vector framework;

applying a projection obtained by learning a low-rank Mahalanobis metric in order to render said first image signature or said second image signature more discriminative; and

matching said first image signature or said second image signature with image signatures stored in said database.