Battery detection method and device

Patent metadata

Field	Value
Publication number	US-11158044-B2
Application number	US-201916650279-A
Country	US
Kind code	B2
Filing date	Jun 27, 2019
Priority date	Aug 27, 2018
Publication date	Oct 26, 2021
Grant date	Oct 26, 2021

Abstract

The present disclosure provides a battery detection method and a battery detection device. The method includes: obtaining a picture of each battery on a battery production line, and obtaining a corresponding production node; inputting the picture into a preset defect detection model, and obtaining a detection result output by the defect detection model, and when the detection result denotes that there is the defect on the picture, sending a control instruction to a control device of the production node corresponding to the picture, to cause the control device to shunt the battery corresponding to the picture having the defect based on the control instruction. The detection result includes whether there is a defect, a defect type, and a defect position.

First claim

What is claimed is:

1. A battery detection method, comprising:
obtaining a picture of each battery on a battery production line, and obtaining a corresponding production node, wherein a plurality of cameras are set at multiple positions on each of a plurality of production nodes of the battery production line to realize the battery detection for components of the battery during production;
inputting the picture into a preset defect detection model, and obtaining a detection result output by the defect detection model, the detection result comprising: whether there is a defect, a defect type, and a defect position;
when the detection result denotes that there is the defect on the picture, sending a control instruction to a control device of the production node corresponding to the picture, to cause the control device to shunt the battery corresponding to the picture having the defect based on the control instruction;
wherein there are a plurality of defect detection models, and the plurality of defect detection models are respectively set on different servers,
inputting the picture into the preset defect detection model and obtaining the detection result output by the defect detection model comprises:
obtaining a load of each of the plurality of defect detection models;
selecting, from the plurality of defect detection models, a first defect detection model whose load satisfies a preset load condition; and
inputting the picture into the first defect detection model to obtain the detection result output by the first defect detection model.

2. The battery detection method of claim 1 , wherein,
the preset defect detection model is a deep neural network model; and
a structure of the preset defect detection model is determined according to a mask Region Convolutional Neural Network RCNN algorithm.

3. The battery detection method of claim 1 , wherein before inputting the picture into the preset defect detection model and obtaining the detection result output by the defect detection model, the battery detection method further comprises:
obtaining training data, the training data comprising historical pictures of batteries on the production line and defect annotation results, and the defect annotation results comprising the defect types and the defect positions;
training an initial defect detection model according to the training data, until a preset loss function satisfies a corresponding condition; and
determining the defect detection model trained as the preset defect detection model.

4. The battery detection method of claim 3 , wherein after inputting the picture into the preset defect detection model and obtaining the detection result output by the defect detection model, the battery detection method further comprises:
reviewing the detection result corresponding to the picture;
adding the picture and the detection result to the training data to obtain updated training data, after the detection result passes the review; and
retraining the defect detection model according to the updated training data.

5. The battery detection method of claim 4 , wherein before retraining the defect detection model according to the updated training data, the battery detection method further comprises:
obtaining a number of pictures and corresponding detection results added into the updated training data;
retraining the defect detection model according to the updated training data comprises:
retraining the defect detection model according to the updated training data when the number is greater than a preset number threshold.

6. A battery detection apparatus, comprising:
a memory;
a processor; and
a computer program stored on the memory and executable by the processor,
wherein when the computer program is executed by the processor, causes the processor to:
obtain a picture of each battery on a battery production line, and obtain a corresponding production node, wherein a plurality of cameras are set at multiple positions on each of a plurality of production nodes of the battery production line to realize the battery detection for components of the battery during production;
input the picture into a preset defect detection model, and obtain a detection result output by the defect detection model, the detection result comprising: whether there is a defect, a defect type, and a defect position;
when the detection result denotes that there is the defect on the picture, send a control instruction to a control device of the production node corresponding to the picture, to cause the control device to shunt the battery corresponding to the picture having the defect based on the control instructions;
wherein there are a plurality of defect detection models, and the plurality of defect detection models are respectively set on different servers,
the processor is further configured to:
obtain a load of each of the plurality of defect detection models;
select, from the plurality of defect detection models, a first defect detection model whose load satisfies a preset load condition; and
input the picture into the first defect detection model to obtain the detection result output by the first defect detection model.

7. A non-transitory computer-readable storage medium having stored thereon a computer program that, when executed by a processor, causes the processor to implement a battery detection method, the method comprising:
obtaining a picture of each battery on a battery production line, and obtaining a corresponding production node, wherein a plurality of cameras are set at multiple positions on each of a plurality of production nodes of the battery production line to realize the battery detection for components of the battery during production;
inputting the picture into a preset defect detection model, and obtaining a detection result output by the defect detection model, the detection result comprising: whether there is a defect, a defect type, and a defect position;
when the detection result denotes that there is the defect on the picture, sending a control instruction to a control device of the production node corresponding to the picture, to cause the control device to shunt the battery corresponding to the picture having the defect based on the control instruction;
wherein there are a plurality of defect detection models, and the plurality of defect detection models are respectively set on different servers,
inputting the picture into the preset defect detection model and obtaining the detection result output by the defect detection model comprises:
obtaining a load of each of the plurality of defect detection models;
selecting, from the plurality of defect detection models, a first defect detection model whose load satisfies a preset load condition; and
inputting the picture into the first defect detection model to obtain the detection result output by the first defect detection model.

8. The battery detection apparatus of claim 6 , wherein,
the preset defect detection model is a deep neural network model; and
a structure of the preset defect detection model is determined according to a mask RCNN algorithm.

9. The battery detection apparatus of claim 6 , wherein the processor is further configured to:
obtain training data, the training data comprising historical pictures of batteries on the production line and defect annotation results, and the defect annotation results comprising the defect types and the defect positions;
train an initial defect detection model according to the training data, until a preset loss function satisfies a corresponding condition; and
determine the defect detection model trained as the preset defect detection model.

10. The battery detection apparatus of

Assignees

• Beijing Baidu Netcom Sci & Tech Co Ltd

Inventors

• Wen Yawei
• Leng Jiabing
• Liu Minghao
• Xiao Huihui
• Guo Jiangliang
• Li Xu

Classifications

• G06T2207/30108

  Industrial image inspection · CPC title

• G06T7/001 Primary

  using an image reference approach · CPC title

• G01N2021/8887

  based on image processing techniques · CPC title

• G01N2021/8809

  Adjustment for highlighting flaws · CPC title

• G01N21/88 Primary

  Investigating the presence of flaws or contamination · CPC title