Vision system

Patent metadata

Field	Value
Publication number	US-10001445-B2
Application number	US-201213628641-A
Country	US
Kind code	B2
Filing date	Sep 27, 2012
Priority date	Sep 27, 2011
Publication date	Jun 19, 2018
Grant date	Jun 19, 2018

Abstract

Provided is a vision system for detecting a defect, including a conveyer belt configured to move at least one bottle through the vision system; a light emitting diode (LED) backlight configured to silhouette any dark contamination on a surface or inside the at least one bottle; a robot with a bottle gripper tooling, the robot configured to pick up the at least one bottle and rotate the at least one bottle; at least one camera configured to take pictures of the at least one bottle while the at least one bottle is being rotated; and a flipper arm configured to stop or allow the at least one bottle to move to a position where the robot picks up the at least one bottle.

First claim

What is claimed is:

1. A method of detecting defects in a vision system, comprising:
obtaining at least one bottle of a material comprising glass, transparent plastic, or non-transparent plastic for determining a defect, wherein determining the defect of the at least one bottle is operational on glass bottles, transparent bottles and non-transparent bottles;
taking a picture of the at least one bottle using at least one camera as the at least one bottle rotates 360 degrees relative to the at least one camera;
inputting the picture;
inputting manual regions that correspond to the imputed picture and excluding regions of the at least one bottle having a grayscale value higher than a predetermined value, each manual region being defined by a plurality of zones, which correspond to the picture taken by the at least one camera;
performing a histogram by calculating grayscale values of each zone;
setting a threshold value of each zone based on the histogram for determining the defect;
comparing the threshold value of each zone with the calculated grayscale values; and
rejecting the at least one bottle if any of the calculated grayscale values is greater than the threshold value of each zone,
wherein the setting threshold value based on each zone based on the histogram further comprises:
calculating a mean grayscale value of each zone based on the grayscale values calculated for the histogram; and
setting the threshold value based on the mean grayscale value of each zone, and

wherein the performing the histogram by calculating grayscale values of each zone further comprises sweeping through each zone and calculating grayscale values at predetermined, non-adjacent interval points of each zone.

2. The method of detecting defects as claimed in claim 1 , wherein the manual regions are defined by a user.

3. The method of detecting defects as claimed in claim 1 , wherein regions having a grayscale value that is less than the predetermined value are included in the manual regions.

4. The method of detecting defects as claimed in claim 1 , wherein the manual regions correspond to the picture by having a same shape as the picture taken by the at least one camera.

5. The method of detecting defects as claimed in claim 1 , further comprising:
stopping the defect detection when the at least one bottle is rejected.

6. The method of detecting defects as claimed in claim 1 , further comprising:
repeating the defect detection for another picture taken by at least one camera when the at least one bottle is not rejected.

7. A method of detecting defects in a vision system, comprising:
obtaining at least one bottle of a material comprising glass, transparent plastic, or non-transparent plastic for determining a defect, wherein determining the defect of the at least one bottle is operational on glass, transparent bottles and non-transparent bottles;
measuring an inner diameter, an outer diameter and threads of the at least one bottle using a neck camera;
rejecting the at least one bottle if the threads or sealing surfaces of the at least one bottle are insufficient based on the measurements of the inner diameter, the outer diameter and the threads;
activating a vacuum of a robot moving the robot toward the at least one bottle to place suction cups on the at least one bottle to lift the at least one bottle of a surface;
taking a picture of the at least one bottle using at least one camera while the robot rotates the at least one bottle 360 degrees relative to the at least one camera with the at least one bottle in an upright position;
inputting the picture;
inputting manual regions that correspond to the inputted picture and excluding regions of the at least one bottle having grayscale value higher than a predetermined value, the manual region comprising a large manual region and a small manual region for sequentially sweeping through the large manual region, the small region corresponding to the large manual region;
performing histogram by calculating grayscale values of each small region;
setting a threshold value of each small region based on the histogram for determining the defect;
comparing the threshold value of each small region with the calculated grayscale values; and
rejecting the at least one bottle if any of the calculated grayscale values is greater than the threshold value of each small region,
wherein the setting the threshold value based on each small manual region based on the histogram further comprises:
calculating a mean grayscale value of each small manual region based on the grayscale values calculated for the histogram; and
setting the threshold value based on the mean grayscale value of each small manual region, and

wherein the performing the histogram by calculating grayscale values of each small region further comprises sweeping each small manual region through the large manual region and calculating grayscale values at predetermined, non-adjacent interval points of each small manual region.

8. The method of detecting defects as claimed in claim 7 , wherein the small manual region corresponds to the large manual region by having a same shape.

9. The method of detecting defects as claimed in claim 7 , wherein the large manual region corresponds to the picture by having a same shape as the picture taken by the at least one camera.

Assignees

• Ring Container Tech
• Ring Container Tech Llc

Inventors

• Childress Jeff
• Kerley Sam
• Vanpelt Danny Wayne

Classifications

• G01N21/9081 Primary

  Inspection especially designed for plastic containers, e.g. preforms · CPC title

• G01N21/909

  in opaque containers or opaque container parts, e.g. cans, tins, caps, labels · CPC title

• G01N21/9009

  Non-optical constructional details affecting optical inspection, e.g. cleaning mechanisms for optical parts, vibration reduction · CPC title