Machine vision-based singulation verification system and method

• US 20020118873A1
• Filed: 01/24/2001
• Published: 08/29/2002
• Est. Priority Date: 01/24/2000
• Status: Active Grant

• Current Assignee: Omron Corporation
• Original Assignee: Robotic Vision Systems, Inc. (Acuity Brands, Inc.)
• Inventors: Tran, Don, Schurr, George W.
• Granted Patent: US 6,944,324 B2
• Time in Patent Office: Days
• US Class Current: 382/143
• CPC Class Codes: G06T 2207/30108 Industrial image inspection
• CPC Class Codes: G06T 7/0004 Industrial image inspection
• CPC Class Codes: G06T 7/13 Edge detection
• CPC Class Codes: G06T 7/181 involving edge growing; involving edge linking
• CPC Class Codes: G06T 7/44 using image operators, e.g. filters, edge density metrics or local histograms

Abstract

A system and method for detecting multiple object conditions, such as side-by-side and overlapped objects, such as packages, on a conveyor. The system includes at least one machine vision system including at least one machine vision camera, at least one illumination subsystem and at least one machine vision computer. The illumination subsystem(s) is configured to illuminate a plurality of objects as they are conveyed past at least one field of view at an inspection station along a conveyor belt. Each machine vision camera is positioned to capture one or more images of the objects as the objects are conveyed past the field of view(s). Each machine vision computer is programmed to detect the presence of multiple object conditions by detecting and counting the number of edges appearing in an image of an object captured by one of the machine vision cameras. The method illuminates at least one object as it passes through the field of view, at which time at least one image of the object is captured. Next, each captured image is processed using a machine vision computer by windowing each parcel using a Region of Interest (ROI) and counting the number of edges appearing within the ROI. The presence of other than a single package condition is determined if the number of edges exceeds four.

Claims

1. A system for detecting the presence of overlapped objects on a conveyor belt conveying a plurality of objects, said system comprising:

at least one machine vision system including at least one machine vision camera, an illumination subsystem and at least one machine vision computer;

said illumination subsystem configured to illuminate said plurality of objects as said objects are conveyed past a field of view at an inspection station along said conveyor belt;

said at least one camera positioned to capture images of said plurality of objects as said objects are conveyed past said field of view, and

said at least one machine vision computer programmed to detect the presence of multiple object conditions by detecting and counting a number of edges appearing in an image captured by said at least one machine vision camera.

2. The system of claim 1, wherein said at least one machine vision camera includes first and second cameras, said first machine vision camera oriented to capture a first image of said objects in said field of view in a vertical direction, and said second machine vision camera oriented to capture a second image of said objects in said field of view in a horizontal direction.

3. The system of claim 2, wherein said illumination subsystem includes first and second illumination sources, said first illumination source for illuminating said objects in a vertical direction and said second illumination source for illuminating said objects in a horizontal direction.

4. The system in claim 3, wherein at least one of said first and second illumination sources includes a lighting system configured to make said objects appear light against a substantially dark background.

5. The system of claim 4, wherein said lighting system includes at least one angled strobe light oriented on a side of said objects coinciding with said at least one machine vision camera.

6. The system of claim 5, wherein said at least one angled strobe includes a xenon strobe.

7. The system of claim 4 wherein said lighting system includes at least one high intensity light adjustably disposed in approximately the same plane as said at least one machine vision camera.

8. The system of claim 7 wherein said at least one high intensity light is a sodium vapor light.

9. The system of claim 7 further comprising a photodiode connected to said lighting system, for controlling the intensity of said high intensity light.

10. The system of claim 4, wherein said lighting system includes a back light oriented on a side of said parcels opposite said at least one machine vision camera.

11. The system of claim 10, wherein said backlight includes an LED (light emitting diode) array.

12. The system of claim 10, wherein said at least one machine vision camera includes a filter sensitive to the bandwidth of said lighting system.

13. The system of claim 1 wherein said objects being conveyed and detected include packages.

16. The method of claim 15 further including the act of identifying parcels as other than cartons by counting blob and hole results from a connectivity analysis.

17. The method of claim 16 further including the act of closing dark holes appearing in an image of at least one parcel identified as a carton caused by artifacts on the carton using a grayscale morphological technique wherein said dark holes are filled with white pixels.

18. The method of claim 15, wherein said act of capturing at least one image of said at least one parcel includes capturing first and second images of at least one parcel, wherein said first image is captured in a vertical direction and said second image is captured in a horizontal image.

19. The method of claim 18, further including the act of determining a gradient angle of said at least one parcel as it appears in said first image and selecting an edge detector to obtain a best outline of said at least one parcel.

20. The method of claim 18, wherein said act of counting a number of edges appearing in said ROI is performed on each of said first and second images.

21. The method of claim 19, wherein said act of counting a number of edges appearing in said ROI includes performing a linear delineation process including a steerable outline edge detection on each windowed parcel image using said gradient angle of each image.

22. The method of claim 21, wherein said steerable outline edge detection includes applying a horizontal kernel and a vertical kernel to said windowed images to emphasize edge contents of substantially horizontal and substantially vertical parcels.

23. The method of claim 21, wherein said steerable outline edge detection includes applying a double kernel edge detector optimized for angles other than horizontal and vertical.

24. The method of claim 15, wherein said act of counting a number of edges appearing in said at least one image includes edge blob sortation, wherein;

a root blob is identified as a blob closest to one corner of said ROI and subsequent blobs are identified and numbered sequentially based on their proximity to a previously identified blob until all blobs are appearing in an image are identified and numbered;

each blob is analyzed to determine if it is a part of an edge and blobs are eliminated using proximity and co-linearity checks;

a number of blobs remaining are counted; and

an overlapping parcel condition is detected if said number of blobs remaining exceeds two in each of the two directions.

14. A system for detecting the presence of overlapped parcels on a conveyor belt conveying a plurality of parcels, said system comprising:

at least one machine vision system including at least one machine vision camera, an illumination subsystem and at least one machine vision computer;

said illumination subsystem configured to illuminate said plurality of parcels as said parcels are conveyed past a field of view at an inspection station along said conveyor belt;

wherein said at least one machine vision camera includes first and second cameras, said first machine vision camera oriented to capture a first image of said objects in said field of view in a vertical direction as said parcels are conveyed past said field of view, and said second machine vision camera oriented to capture a second image of said objects in said field of view in a horizontal direction as said parcels are conveyed past said field of view, and

said at least one machine vision computer programmed to detect the presence of multiple parcel conditions by detecting and counting a number of edges appearing in an image of an parcel captured by said at least one machine vision camera.

15. A method of detecting the presence of overlapping parcels on a conveyor belt conveying a plurality of parcels past an inspection station including a machine vision system having at least one machine vision camera for capturing images within a field of view, an illumination subsystem for illuminating said parcels as said parcels are conveyed through said field of view on said conveyor belt and a machine vision computer, said overlapping parcel detection method including the acts of:

illuminating at least one parcel as said at least one parcel passes through said field of view;

capturing at least one image of said at least one parcel as said at least one parcel is illuminated;

processing said at least one captured image using said machine vision computer by windowing said at least one parcel using a Region of Interest (ROI), counting a number of edges appearing in said ROI, and determining the presence of other than a single carton if said number of edges counted exceeds four.