Produce recognition method

• US 6,577,983 B1
• Filed: 10/06/2000
• Issued: 06/10/2003
• Est. Priority Date: 10/06/2000
• Status: Expired due to Term

• Current Assignee: NCR Voyix Corporation
• Original Assignee: NCR Corporation (NCR Voyix Corporation)
• Inventors: Zhu, Jie
• Primary Examiner(s): Barlow, John
• Assistant Examiner(s): LAU, TUNG S
• Application Number:
  US09/684,414
• Time in Patent Office: 977 Days
• Field of Search: 702/128, 235/463, 342/162, 395/135, 381/42, 233/462, 177/25, 177/50, 356/419, 356/416, 356/420, 356/226, 356/328, 250/226, 382/190, 382/110
• US Class Current: 702/128
• CPC Class Codes: G01N 2021/0389 Windows
• CPC Class Codes: G01N 21/255 Details, e.g. use of specially adapted sources, lighting or optical systems
• CPC Class Codes: G01N 21/31 Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry G01N21/72 takes precedence
• CPC Class Codes: G01N 21/55 Specular reflectivity
• CPC Class Codes: G01N 2201/0616 Ambient light is used
• CPC Class Codes: G01N 2201/0622 Use of a compensation LED
• CPC Class Codes: G01N 2201/0624 Compensating variation in output of LED source
• CPC Class Codes: G01N 2201/0627 Use of several LED's for spectral resolution
• CPC Class Codes: G01N 2201/1244 Ambient light detector, e.g. for invalidating
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: G06V 20/68 Food, e.g. fruit or vegetables
• 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: H04N 1/32101 Display, printing, storage or transmission of additional information, e.g. ID code, date and time or title
• CPC Class Codes: H04N 2201/3226 of identification information or the like, e.g. ID code, index, title, part of an image, reduced-size image
• CPC Class Codes: H04N 2201/3273 Display
• CPC Class Codes: H04N 2201/3274 Storage or retrieval of prestored additional information

Abstract

A produce recognition method which determines an optimal number of candidate identifications in a candidate identification list. The method includes the steps of obtaining produce data associated with a produce item, determining distances between the produce data and reference produce data, determining confidence values from the distances, determining first confidence values which are greater than a threshold confidence value, displaying candidate identifications associated with the first confidence values, and recording an operator choice of one of the candidate identifications.

Claims

1. A produce recognition system for recognizing a non-bar coded produce item comprising:

a storage medium having a produce database and a look up data file, the produce database stores reference produce data items and item identification numbers;

a produce data collector for receiving light data reflected from the non-bar coded produce item and converting the light data to a digitized collected produce data;

wherein the produce data collector includes

a housing;

a window mounted within the housing; and

a light source having one or more light emitted diodes for producing a white light spectral distribution, the white light transmitted through the window on to the non-bar coded produce item to create a reflected light;

spectrometer having a light-separating element and a photodetector array, the light separating element converts the reflected light received through the window to a continuous band of wavelengths, the photodetector array spatially samples the continuous band of wavelengths to produce discrete signal levels; and

a control circuit converts the discrete signal levels to digitized produce data;

a transaction software component;

a processor and produce recognition software for comparing the digitized collected produce data with the reference produce data items, retrieving the item identification number for the non-bar coded produce item on a successful match of the digitized collected produce data and the reference produce data item, and obtaining a corresponding price for the non-bar coded item from the look up data file; and

a display for displaying the price retrieved from the look up data file.

3. A produce recognition system as in claim 1 wherein the produce data collector further comprises:

an ambient light sensor mounted within the housing and in direct view of the window, the sensor receives ambient light through the window and sends signals to the control circuit to begin transmitting light data at the produce item.

2. A produce recognition system for recognizing a non-bar coded produce item comprising:

a storage medium having a produce database, the produce database stores reference produce data items;

a produce data collector for receiving light data reflected from the non-bar coded produce item and converting the light data to a digitized collected produce data;

wherein the produce data collector includes

a housing;

a window mounted within the housing;

a light source having one or more light emitted diodes for producing a white light spectral distribution, the white light transmitted through the window on to the non-bar coded produce item to create a reflected light;

a spectrometer having a light separating element and a photodetector array, the light separating element converts the reflected light received through the window to a continuous band of wavelengths, the photodetector array spatially samples the continuous band of wavelengths to produce discrete signal levels; and

a control circuit converts the discrete signal levels to digitized produce data;

a transaction software component;

a processor and produce recognition software comparing the digitized collected produce data with reference produce data items to define a candidate list of reference data describing like produce, the candidate list having a confidence interval for each chosen reference produce data item and sorted thereby, each chosen reference produce data item is chosen by a distance measuring algorithm, the list size determined by summing the confidence values to a value greater than or equal to a threshold probability representing success; and

a display for displaying the candidate list and allowing the operator to choose from the list.

4. A produce recognition system as in claim 2 wherein the produce data collector further comprises:

an ambient light sensor mounted within the housing and in direct view of the window, the sensor receives ambient light through the window and sends signals to the control circuit to begin transmitting light data at the produce item.

5. A non-bar coded produce recognition method comprising the steps of:

(a) receiving light spectral data describing the produce item;

(b) converting the light spectral data to digitized collected produce data;

(c) comparing the digitized collected produce data with the reference produce data stored in the produce database, on a successful match an item identification number associated with the reference produce data is passed to a transaction software component, the transaction software component obtaining a corresponding unit price from a look up data file;

(d) obtaining weight information associated with the produce item; and

(e) calculating the overall produce data cost by multiplying the unit price by weight information.

6. A non-bar coded produce recognition method comprising the steps of:

(a) receiving light spectral data reflected off of the produce item;

(b) converting the light spectral data to a digitized collected produce data;

(c) comparing the digitized collected produce data with reference produce data items stored in a produce database to define a sorted candidate list of reference data items, the candidate list having a confidence interval for each chosen reference produce data item within the list, each chosen reference produce data chosen by a distance measuring algorithm, the list size determined by summing the confidence values to a value greater than or equal to a threshold probability representing success;

(d) displaying the list;

(e) recording an operator choice for a selected item from a input device;

(f) obtaining a unit price for the selected item;

(g) obtaining weight information associated with the produce from a scale; and

(h) calculating the overall produce data cost by multiplying the unit price by the weight information.

7. A non-bar coded produce recognition method comprising the steps of:

(a) receiving light spectral data reflected off of the produce item;

(b) converting the light spectral data to a digitized collected produce data;

(c) comparing the digitized collected produce data with reference produce data items stored in a produce database to define a sorted candidate list of reference data items, the candidate list having a confidence interval for each chosen reference produce data item within the list, each chosen reference produce data chosen by a distance measuring algorithm, the list size determined by summing the confidence values to a value greater than or equal to a threshold probability representing success; and

(d) displaying the list and images associated with each entry in the list.