System and methods for preventing fraud in retail environments, including the detection of empty and non-empty shopping carts

• US 5,883,968 A
• Filed: 02/09/1996
• Issued: 03/16/1999
• Est. Priority Date: 07/05/1994
• Status: Expired due to Fees

• Current Assignee: AW Computer Systems, Inc.
• Original Assignee: AW Computer Systems, Inc.
• Inventors: Rozmus, J. Michael, Negin, Michael, Herd, William J., Whiteman, James L., Welch, Charles W.
• Primary Examiner(s): CHANG, JON CARLTON
• Application Number:
  US08/598,395
• Time in Patent Office: 1,131 Days
• Field of Search: 382/165, 382/170, 382/141-143, 382/100, 382/147, 382/160, 382/164, 382/181, 382/191, 382/128, 382/133, 358/522, 358/520, 235/383, 235/454, 235/462, 235/467, 186/55, 186/61, 186/62, 177/50
• US Class Current: 382/100
• CPC Class Codes: A47F 9/045 Handling of baskets or shopping trolleys at check-out counters, e.g. unloading, checking
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 2207/30108 Industrial image inspection
• CPC Class Codes: G06T 7/001 using an image reference approach
• CPC Class Codes: G07G 1/0036 Checkout procedures
• CPC Class Codes: G07G 1/0054 with control of supplementary check-parameters, e.g. weight or number of articles
• CPC Class Codes: G07G 3/003 Anti-theft control theft alarms G08B13/00

Abstract

Purchase of an item bearing a unique identification code is verified by (a) ascertaining the identification code of an item to be purchased; (b) acquiring data representative of color features of the item to be purchased; and (c) comparing the data representative of color features of the item to be purchased to data representative of expected color features for items having the same identification code. Other physical characteristics of an item, such as height, may also be employed. The disclosed system is also capable of detecting whether a shopping cart is empty or not, and determining whether a non-empty cart is an indication that a theft is being attempted.

Claims

1. A method for use in preventing fraud in a retail or like environment in which customers utilize carts to bring items to be purchased to a checkout station having a cart aisle through which the carts pass, said method comprising the steps of:

(a) providing empty cart image data representing at least one image of a cart that is known to be empty;

(b) when a customer arrives at said checkout station with a cart to perform a transaction for the purchase of selected items, obtaining an image of said customer cart;

(c) comparing the image of said customer cart to said empty cart image data, and determining from said comparison whether said customer cart is empty;

wherein said items to be purchased have identifying codes associated with them, said method further comprising the steps of;

detecting the codes associated with items to be purchased in said transaction as said items are removed from said customer cart;

if said cart is determined not to be empty, determining whether a code has been detected for an item that is permitted to be in the cart; and

if no code is detected for an item that is permitted to be in the cart, initiating an action to deter a possible attempt by the customer to leave the premises without paying for an item in the cart.

2. The method recited in claim 1 further comprising the steps of:

(i) if said customer cart is determined not to be empty, determining whether the non-empty cart is permitted for said transaction; and

(ii) if said non-empty cart is not permitted for said transaction, initiating an action to deter a possible attempt by the customer to leave the premises without paying for an item in the cart.

3. The method recited in claim 1 wherein an optically discernable strip is affixed at a predetermined location on said customer cart, and wherein the step of obtaining an image of said customer cart comprises the steps of:

obtaining an image of an area of said cart aisle adjacent said checkout station;

processing the image of said cart aisle until said optically discernable strip is detected within said image of the cart aisle;

defining, based on the location of said optically discernable strip in the image of said cart aisle, an area of interest in said image corresponding to at least a portion of said customer cart, said area of interest defining said image of said customer cart.

4. The method recited in claim 3 wherein said optically discernable strip has a pre-defined color, and wherein said step of processing the image of said cart aisle to detect said optically discernable strip comprises the steps of:

scanning the image of said cart aisle to detect points of transition in the image between the pre-defined color of said strip and other colors in the image; and

processing the detected points of transition in accordance with a least-squares-fit algorithm to generate a line representative of the location of said strip.

5. The method recited in claim 3 wherein first and second optically discernable strips are attached to a front end and a rear end, respectively, of said customer cart and wherein said area of interest in the image of the cart aisle is defined based on the location of either one of said first and second strips within the image.

6. The method recited in claim 1 wherein the empty cart image data comprises a histogram representing color features of at least one image of a known empty cart and wherein the image of said customer cart is compared to the empty cart image data by generating a histogram representing color features of the image of said customer cart, and then comparing that histogram to the empty cart image data.

7. The method recited in claim 6 wherein the image of said customer cart comprises a plurality of pixels at least some of which having a determinable hue, and wherein the histogram representing color features of the image of said customer cart is derived from the hue of different pixels in the image.

8. The method recited in claim 7 wherein the histogram representing color features of the image of said customer cart has a plurality of bins, each bin representing a different range of hues and containing a count of the number of pixels in the image of said customer cart that have a hue within that range.

9. The method recited in claim 6 wherein said empty cart image data comprises a statistical representation of color feature data generated from a plurality of images of at least one known empty cart.

10. A security system for use in preventing fraud in a retail or like environment in which customers utilize carts to bring items to be purchased to a checkout station having a cart aisle through which the carts pass, said system comprising:

(a) means for obtaining and storing first data representative of an image of a known empty cart;

(b) means for obtaining second data representative of an image of a customer cart when a customer arrives at said checkout station with said customer cart to perform a transaction for the purchase of selected items; and

(c) means for comparing said first and second data and for determining from said comparison whether said customer cart is empty;

wherein said items to be purchased in said transaction have identifying codes associated with them, and wherein the system further comprises;

means for detecting the codes associated with items to be purchased in said transaction as said items are removed from said customer cart;

means for determining whether a code has been detected for an item that is permitted to be in the customer cart, if said customer cart is determined not to be empty; and

means for initiating an action to deter a possible attempt by the customer to leave the premises without paying for an item in the customer cart, when said cart is determined not to be empty and no code has been detected for an item that is permitted to be in the cart.

11. The system recited in claim 10 further comprising means for initiating an action to deter a possible attempt by the customer to leave the premises without paying for an item in the customer cart, if it is determined that the customer cart is not empty and that a non-empty cart is not permitted for said transaction.

12. The system recited in claim 10 further comprising:

an optically discernable strip affixed at a predetermined location on said customer cart;

means for obtaining an image of an area of said cart aisle adjacent said checkout station and for processing the image of said cart aisle until said optically discernable strip is detected within the image; and

means for defining, based on the location of said optically discernable strip within the image of said cart aisle, an area of interest in the image corresponding to at least a portion of said customer cart, said area of interest defining the image of said customer cart.

13. The system recited in claim 12 wherein said optically discernable strip has a pre-defined color, and wherein said means for processing the image of said cart aisle comprises:

means for scanning the image of said cart aisle to detect points of transition in the image between the pre-defined color of said strip and other colors in the image; and

means for processing the detected points of transition in accordance with a least-squares-fit algorithm to generate a line representative of the location of said strip.

14. The system recited in claim 12 comprising first and second optically discernable strips attached to a front end and a rear end, respectively, of said customer cart.

15. The system recited in claim 10 wherein said first data comprises a histogram representing color features of at least one image of a known empty cart and wherein said second data comprises a histogram representing color features of the image of said customer cart.

16. The system recited in claim 15 wherein the image of said customer cart comprises a plurality of pixels at least some of which having a determinable hue, and wherein the histogram representing color features of the image is derived from the hue of different pixels therein.

17. The system recited in claim 16 wherein the histogram representing color features of the image of said customer cart has a plurality of bins, each bin representing a different range of hues and containing a count of the number of pixels in the image that have a hue within that range.

18. The system recited in claim 15 wherein the first data comprises a statistical representation of color features of a plurality of images of at least one known empty cart.