Systems and Methods for Inspecting Absorbent Articles on A Converting Line

• US 20180357756A1
• Filed: 05/31/2018
• Published: 12/13/2018
• Est. Priority Date: 06/13/2017
• Status: Active Grant

• Current Assignee: The Procter & Gamble Company
• Original Assignee: The Procter & Gamble Company
• Inventors: Kawka, Paul Anthony, Varga, Stephen Michael, Ahmad, Aitzaz, Thomas, Paul Philip
• Granted Patent: US 10,632,023 B2
• Time in Patent Office: Days
• CPC Class Codes: A61F 13/15585 of babies' napkins, e.g. diapers A61F13/15617 - A61F13/15804 take precedence
• CPC Class Codes: A61F 13/15772 Control
• CPC Class Codes: A61F 13/49 Absorbent articles specially adapted to be worn around the waist, e.g. diapers
• CPC Class Codes: A61F 13/49009 with elastic means
• CPC Class Codes: A61F 13/495 with faecal cavity
• CPC Class Codes: A61F 2013/1578 via optical means associated to computer coordinator
• CPC Class Codes: A61F 2013/15788 of the presence of the article or components
• CPC Class Codes: A61F 2013/15796 of the alignment or position of article or components
• CPC Class Codes: B32B 5/26 another layer next to it also being fibrous or filamentary relative arrangement of fibres or filaments of different layers B32B5/12
• CPC Class Codes: B32B 5/32 at least two layers being foamed and next to each other
• CPC Class Codes: D04H 1/593 to layered webs
• CPC Class Codes: G06N 3/045 Combinations of networks
• CPC Class Codes: G06N 3/08 Learning methods
• CPC Class Codes: G06N 5/046 Forward inferencing; Production systems
• CPC Class Codes: G06T 2207/20084 Artificial neural networks [ANN]
• CPC Class Codes: G06T 2207/30108 Industrial image inspection
• CPC Class Codes: G06T 7/0002 Inspection of images, e.g. flaw detection
• CPC Class Codes: G06T 7/0004 Industrial image inspection
• CPC Class Codes: G06T 7/11 Region-based segmentation
• CPC Class Codes: G06T 7/187 involving region growing; involving region merging; involving connected component labelling
• CPC Class Codes: G06T 7/97 : Determining parameters from multiple pictures depth or shape recovery from multiple images G06T7/55; stereo camera calibration G06T7/85

Abstract

A method for inspecting absorbent articles is provided. The inspection is performed using an inspection algorithm generated with a convolutional neural network having convolutional neural network parameters. The convolutional neural network parameters are generated by a training algorithm. Based on the inspection, characteristics of the absorbent articles, such as defects, can be identified. Absorbent articles having identified characteristics can be rejected, or other actions can be taken.

Claims

1. A method for inspecting absorbent articles, the method comprising the steps of:

preprocessing each image of a group of images of absorbent articles to label a component of the absorbent article;

generating a first inspection algorithm with a convolutional neural network based on the preprocessed images, wherein the first inspection algorithm is usable to mask the labeled component in images of absorbent articles;

providing a second inspection algorithm, wherein the second inspection algorithm is usable to determine one or more properties of the labeled component in images of absorbent articles, wherein the one or more properties comprise any of a location, a size, a shape, and a presence of the component;

providing a communication network;

connecting a sensor with the communication network;

connecting a controller with the communication network, the controller comprising the first inspection algorithm;

advancing a substrate through a converting process;

sequentially adding component parts to the substrate;

creating images of at least one of the substrate and component parts with the sensor;

communicating the images from the sensor to the controller;

masking the labeled component in the images by analyzing the images with the first inspection algorithm;

determining at least one of the properties for the labeled component in the substrate with the second inspection algorithm;

cutting the substrate with component parts added thereto into discrete absorbent articles; and

based on the determination of the least one of the properties for the labeled component in the substrate, executing a control action, wherein the control action is any of rejecting one or more of the discrete absorbent articles, an automatic phase adjustment, an automatic tracking adjustment, a machine maintenance scheduling, and a machine stop command.

2. The method of claim 1, wherein the controller comprises any of a field programmable gate array, a graphics processing unit, and a central processing unit.

3. The method of claim 1, wherein the component parts include parts added as a continuous web of material and parts added as a discontinuous web of material.

4. The method of claim 1, wherein the communication network comprises any of an Ethernet protocol, a USB protocol, a FireWire protocol, and a CameraLink protocol.

5. The method of claim 1, wherein the labeled component comprises any of a contaminant and an undesirable transformation of the absorbent article.

6. The method of claim 5, wherein the undesirable transformation of the absorbent article comprises any of a hole, a tear, and a wrinkle.

7. The method of claim 1, wherein determining the at least one of the properties for the labeled component in the substrate with the second inspection algorithm comprises determining an angular orientation of the labeled component.

8. The method of claim 1, wherein determining the at least one of the properties for the labeled component in the substrate with the second inspection algorithm comprises determining the absence of the labeled component.

9. The method of claim 1, wherein determining the at least one of the properties for the labeled component in the substrate with the second inspection algorithm comprises determining a texture of the labeled component.

10. The method of claim 1, wherein determining the at least one of the properties for the labeled component in the substrate with the second inspection algorithm comprises determining a total number of the labeled components.

11. The method of claim 1, wherein determining the at least one of the properties for the labeled component in the substrate with the second inspection algorithm comprises determining a deformation of the labeled component.

12. The method of claim 1, wherein the absorbent articles are any of diapers and feminine hygiene products.

13. The method of claim 12, wherein the component parts comprise any of ears, tapes, graphics, elastics, core, cuts, and wrappers.

14. A method for inspecting absorbent articles, the method comprising the steps of:

preprocessing each image of a group of images of absorbent articles to label a component of the absorbent article;

generating a first inspection algorithm with a convolutional neural network based on the preprocessed images, wherein the first inspection algorithm is usable to mask the labeled component in images of absorbent articles;

providing a second inspection algorithm, wherein the second inspection algorithm is usable to determine one or more properties of the labeled component in images of absorbent articles, wherein the one or more properties comprise any of a location, a size, a shape, and a presence of the component;

advancing a substrate through a converting process;

sequentially adding component parts to the substrate;

creating images of a substrate and one or more component parts;

masking the labeled component in the images by analyzing the images with the first inspection algorithm; and

determining at least one of the properties for the labeled component in the substrate with the second inspection algorithm.

15. The method of claim 14, further comprising:

cutting the substrate with component parts added thereto into discrete absorbent articles; and

based on the determination of the least one of the properties for the labeled component in the substrate, executing a control action.

16. The method of claim 15, wherein the control action is any of rejecting one or more of the discrete absorbent articles, an automatic phase adjustment, an automatic tracking adjustment, a machine maintenance scheduling, and a machine stop command.

17. The method of claim 14, wherein determining the at least one of the properties for the labeled component in the substrate with the second inspection algorithm comprises any of determining an angular orientation of the labeled component, determining the absence of the labeled component, determining a texture of the labeled component, determining a total number of the labeled components, determining the presence of a contaminant, and determining an undesirable transformation of the labeled component.

18. A method for inspecting absorbent articles, the method comprising the steps of:

storing a group of digitized signals of absorbent articles, wherein the digitized signals are preprocessed to label a component of the absorbent article;

providing the preprocessed digitized signals to a training algorithm to create a convolutional neural network;

generating a first inspection algorithm with the convolutional neural network based on the group of preprocessed digitized signals, wherein the first inspection algorithm masks one or more properties of the labeled component;

providing a second inspection algorithm, wherein the second inspection algorithm is usable to determine one or more properties of the labeled component in digitized signals of absorbent articles, wherein the one or more properties comprise any of a location, a size, a shape, and a presence of the component;

advancing a substrate through a converting process;

sequentially adding component parts to the substrate;

creating inspection data of at least one of the substrate and component parts with a sensor;

communicating the inspection data from the sensor to a controller, wherein the controller comprises the first inspection algorithm;

identifying the labeled component in the substrate by analyzing the inspection data with the first inspection algorithm;

determining at least one of the properties for the labeled component in the substrate with the second inspection algorithm; and

cutting the substrate with component parts added thereto into discrete absorbent articles.

19. The method of claim 18, further comprising the step of executing a control action based on the determination of the least one of the properties for the labeled component in the substrate, wherein the control action is selected from the group consisting of:

adjusting an advancement speed of the substrate;

adjusting a placement of component parts; and

rejecting absorbent articles.

20. The method of claim 18, wherein the digitized signals are selected from the group consisting of:

images;

profile signals;

3-D point cloud;

height maps; and

sensor data.