Distributed computing system for product defect analysis

Patent metadata

Field	Value
Publication number	US-11880968-B2
Application number	US-201917043960-A
Country	US
Kind code	B2
Filing date	Dec 20, 2019
Priority date	Dec 20, 2019
Publication date	Jan 23, 2024
Grant date	Jan 23, 2024

Abstract

A distributed computing system for product defect analysis is disclosed. The distributed computing system for product defect analysis includes a computing cluster for processing product manufacturing messages, a computing cluster for identifying product defect, a product image database, and a client device.

First claim

The invention claimed is:

1. A distributed computing system for product defect analysis, comprising a computing cluster for processing product manufacturing messages, a computing cluster for identifying product defect, a product image database, and a client device, wherein:
the product image database is configured to store product images generated in the product manufacturing process;
the computing cluster for processing product manufacturing messages is configured to generate product defect analysis tasks according to product manufacturing messages, wherein the product defect analysis tasks include a first defect identification task;
the computing cluster for identifying product defect is configured to identify product defect content in one of the product images through a defect identification model based on the first defect identification task, wherein the product defect content includes any one or more of: defect type, defect position, and defect size of products; and
the client device is configured to output a product defect identification result,
wherein the product manufacturing messages include a first product manufacturing message related to registration information, and the computing cluster for processing product manufacturing messages is configured to monitor the first product manufacturing message by interrupt,
wherein the product manufacturing messages further include a second product manufacturing message irrelevant to the registration information; and the computing cluster for processing product manufacturing messages is further configured to: monitor the second product manufacturing message by polling, determine whether the second product manufacturing message includes a product manufacturing keyword; keep the second product manufacturing message if the second product manufacturing message includes the product manufacturing keyword; and discard the second product manufacturing message if the second product manufacturing message does not include the product manufacturing keyword.

2. The distributed computing system for product defect analysis according to claim 1 , wherein the first product manufacturing message includes a lot-product manufacturing message, and the second product manufacturing message includes a single product manufacturing message.

3. The distributed computing system for product defect analysis according to claim 1 , wherein generating the product defect analysis tasks according to the product manufacturing messages includes:
determining whether a defect identification model corresponding to the product type exists according to the product manufacturing message;
generating the first defect identification task if a defect identification model corresponding to the product type exists; and
sending a first defect analysis task to the computing cluster for identifying product defect.

4. The distributed computing system for product defect analysis according to claim 3 , wherein generating the product defect analysis tasks according to the product manufacturing messages further includes:
generating a second defect identification task if no defect identification model is corresponding to the product type; and
sending a second defect analysis task to the client device.

5. The distributed computing system for product defect analysis according to claim 1 , wherein the computing cluster for processing product manufacturing messages is configured to:
establish a product defect analysis task queue by sorting a plurality of product defect analysis tasks based on one or more of: the time sequence of generating the plurality of product defect analysis tasks, a priority order of products, and a product scheduling plan.

6. The distributed computing system for product defect analysis according to claim 5 , wherein the computing cluster for processing product manufacturing messages is further configured to:
control a speed of sending the product defect analysis tasks in the product defect analysis task queue according to a load status of the computing cluster for identifying product defect.

7. The distributed computing system for product defect analysis according to claim 1 , wherein the defect identification model includes any one or more of: a feedforward neural network defect identification model, a convolutional neural network model, a recurrent neural network model, and a generative adversarial network model.

8. The distributed computing system for product defect analysis according to claim 1 , wherein the client device includes a display; and wherein outputting the product defect identification result further includes:
displaying, by the display, a product image of which product defect is identified by the computing cluster for identifying product defect.

9. The distributed computing system for product defect analysis according to claim 8 , wherein the product image of which product defect is identified by the computing cluster for identifying product defect is labeled with a product defect content label, and the product defect content label includes any one or more of: defect mark box, defect mark box coordinate, defect category label, and defect category score.

10. The distributed computing system for product defect analysis according to claim 8 , wherein the display is configured to display a product image of which the product defect cannot be identified by the computing cluster for identifying product defect.

11. The distributed computing system for product defect analysis according to claim 10 , wherein the client device is also configured to:
obtain a product defect identification instruction inputted by a user; and
display the product defect content label of the product images of which the product defect cannot be identified by the computing cluster for identifying product defect based on the product defect identification instruction.

12. The distributed computing system for product defect analysis according to claim 9 , wherein the client device is further configured to:
obtain a product defect rechecking instruction inputted by a user; and
confirm or modify the product defect content label in a product image to be rechecked based on the product defect rechecking instruction.

13. The distributed computing system for product defect analysis according to claim 8 ,
wherein the display is further configured to:
display a task list corresponding to a plurality of product defect analysis tasks;
wherein the client device is further configured to:
obtain an execution priority adjustment instruction of the plurality of product defect analysis tasks inputted by the user; and
modify the sorting of the product defect analysis tasks in the task list based on the execution priority adjustment instruction.

14. The distributed computing system for product defect analysis according to claim 1 , wherein
the computing cluster for processing product manufacturing messages is further configured to generate a defect identification model training task; and
the computing cluster for identifying product defect is further configured to train the defect identification model based on the defect identification model training task.

15. A product defect analysis method based on a distributed system, comprising:
generating product defect analysis tasks according to product manufacturing messages, wherein the product defect analysis tasks include a first defect identification task;
identifying product defect content in a product image by a defect identification model based on the first defect identification task, wherein the product defect content includes any one or more of: defect type, def

Assignees

• Boe Technology Group Co Ltd

Inventors

• Li Zhaoyue
• Chai Dong
• Wang Yaoping
• Zhang Meijuan
• Wang Hong

Classifications

• Y02P90/02

  Total factory control, e.g. smart factories, flexible manufacturing systems [FMS] or integrated manufacturing systems [IMS] · CPC title

• G05B2219/32368

  Quality control · CPC title

• G06Q10/06395

  Quality analysis or management · CPC title

• G06N3/0475

  Generative networks · CPC title

• G06N3/044

  Recurrent networks, e.g. Hopfield networks · CPC title