Optical vibrometric testing of container for items

Patent metadata

Field	Value
Publication number	US-10229487-B2
Application number	US-201715443677-A
Country	US
Kind code	B2
Filing date	Feb 27, 2017
Priority date	Feb 27, 2017
Publication date	Mar 12, 2019
Grant date	Mar 12, 2019

Abstract

Techniques for automated quality control of containers and items are disclosed. Images of a container can be successively captured over time. A consolidated image can be generated from the captured images. A non-image representation of the consolidated image can be determined. The non-image representation can be used to determine whether the container satisfies a condition. An imaging system can include a visual reference object or an object sensor used to detect entry of the container into a view volume of an electronic camera. A motion system can transport the container into the view volume. Some examples operate in an automated-warehouse environment.

First claim

What is claimed is:

1. A method of verifying contents of a container, the method comprising:
successively capturing multiple stripe images using an electronic camera operating at a frame rate of at least 60 frames per second, the electronic camera including a common capture region that images at least part of the container;
generating a consolidated image based at least in part on the multiple stripe images;
determining a non-image representation of the consolidated image;
determining a reference non-image representation based at least in part on a type of the container, wherein the reference non-image representation is associated with reference contents of the container;
comparing the non-image representation with the reference non-image representation to provide an indication of similarity; and
determining whether contents of the container correspond to the reference contents of the container based at least in part on the indication of similarity.

2. The method of claim 1 , wherein:
the reference contents comprise a predetermined number of items; and
the method further comprises, in response to determining that the contents of the container do not correspond to the reference contents of the container, providing an alert indicating that the container has an incorrect number of items.

3. The method of claim 1 , further comprising:
detecting entry of the container at least partly into a view volume associated with the common capture region; and
subsequently, capturing the multiple stripe images as consecutive frames provided by the electronic camera.

4. A system, comprising:
an electronic image sensor;
one or more processors, at least one of which is communicatively coupled to the electronic image sensor; and
memory coupled to the one or more processors, the memory including instructions that, when executed by the one or more processors, cause the one or more processors to perform acts comprising:
successively capturing individual images of a plurality of images using the electronic image sensor;
generating a consolidated image based at least in part on the plurality of images;
determining a non-image representation of the consolidated image;
determining whether the non-image representation satisfies a predetermined criterion; and
providing an alert in response to the non-image representation failing to satisfy the predetermined criterion.

5. The system of claim 4 , wherein:
at least one of the one or more processors is communicatively connectable with a motion system; and
the acts comprise, before capturing the individual images, causing the motion system to position an item at least partly within a view volume of the electronic image sensor.

6. The system of claim 4 , wherein:
the electronic image sensor comprises pixels of a common capture region and pixels outside the common capture region;
the system comprises optics arranged with respect to the electronic image sensor to define a view volume of the common capture region; and
the acts comprise capturing the individual images using the pixels of the common capture region.

7. The system of claim 4 , further comprising a second electronic image sensor, wherein the acts comprise:
successively capturing second individual images of a second plurality of images using the second electronic image sensor;
generating a second consolidated image based at least in part on the second plurality of images;
determining a second non-image representation of the second consolidated image; and
determining whether the second non-image representation satisfies a second predetermined criterion.

8. The system of claim 4 , wherein the acts comprise:
detecting entry of a container at least partly into a view volume of the electronic image sensor; and
subsequently, capturing the individual images.

9. The system of claim 8 , further comprising an object sensor configured to detect the container in the view volume of the electronic image sensor, wherein the acts comprise detecting the entry of the container using the object sensor.

10. The system of claim 4 , further comprising a visual reference object arranged at least partly within a view volume of the electronic image sensor, wherein the acts comprise:
successively capturing a plurality of calibration images of the visual reference object using the electronic image sensor;
determining a vibration profile associated with the electronic image sensor based at least in part on the plurality of calibration images; and
determining whether the non-image representation satisfies the predetermined criterion further based at least in part on the vibration profile.

11. The system of claim 4 , wherein:
the electronic image sensor comprises pixels of a common capture region and pixels outside the common capture region;
the pixels of the common capture region are arranged substantially along a primary capture axis;
the system comprises a motorized mount retaining the electronic image sensor; and
the acts comprise:
determining a target orientation of a target object at least partly within a view volume of the electronic image sensor; and
causing the motorized mount to orient the electronic image sensor based at least in part on the target orientation so that the primary capture axis is aligned either substantially parallel to or substantially transverse to the target orientation.

12. A method comprising:
generating a consolidated image based at least in part on a plurality of successively-captured images depicting at least a portion of a test article that includes at least one container;
determining a reference image based in part on at least one of the container or the test article;
comparing the consolidated image with the reference image to provide an indication of similarity; and
determining whether the consolidated image satisfies a predetermined criterion associated with a condition of the test article based at least in part on the indication of similarity.

13. The method of claim 12 , further comprising:
determining a non-image representation of the consolidated image based at least in part on a basis, wherein the basis is associated with the condition; and
determining whether the non-image representation satisfies the predetermined criterion.

14. The method of claim 13 , wherein the condition comprises a first sub-condition and a second sub-condition, the method further comprising determining the basis based at least in part on first basis data associated with the first sub-condition and second basis data associated with the second sub-condition.

15. The method of claim 13 , further comprising determining the basis by performing an eigenvalue decomposition of a plurality of training images associated with the condition.

16. The method of claim 12 , further comprising operating a computational model based at least in part on the consolidated image to provide data indicating whether the consolidated image satisfies the predetermined criterion, wherein the computational model is associated with a training data set comprising a plurality of training consolidated images.

17. The method of claim 16 , wherein the computational model comprises at least one of a neural network, a convolutional neural network, a decision tree, a random forest, a Bayesian classifier, or a support vector machine (SVM).

18. The method of claim 12 , further comprising successively capturing the individual images of the plurality of images using an electronic image sensor, wherein at least a portion of the container or of an item in the container

Assignees

• Amazon Tech Inc

Inventors

• Goyal Dushyant
• Mishra Pragyana K

Classifications

• G06Q10/087 Primary

  Inventory or stock management, e.g. order filling, procurement or balancing against orders · CPC title

• G06T7/0008 Primary

  checking presence/absence · CPC title

• G06K9/6202

  Physics · mapped topic

• G06T2207/30242

  Counting objects in image · CPC title

• G06T2207/10016

  Video; Image sequence · CPC title