Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Sealing ring classification model training method, sealing ring classification method and device

A classification model and training method technology, which is applied in the training field of sealing ring classification model, can solve the problems of detection result error and low detection accuracy

Pending Publication Date: 2021-03-05
WATRIX TECH CORP LTD +1
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, when this method is applied to non-rigid O-rings, such as rubber O-rings, the distance from the edge of the O-ring to the center of the O-ring will change due to the deformation of the rubber O-ring. Judging by the above method When the rubber O-ring is a good product, it is easy to judge the good product O-ring that is deformed but has no burrs as a non-good product, which makes the detection result have large errors and the detection accuracy is low.

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Sealing ring classification model training method, sealing ring classification method and device
  • Sealing ring classification model training method, sealing ring classification method and device
  • Sealing ring classification model training method, sealing ring classification method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0056] The embodiment of the present application provides a kind of training method of sealing ring classification model, such as figure 1 A flowchart of a training method for a sealing ring classification model shown, wherein the sealing ring classification model includes a plurality of convolutional neural networks and classifiers; the training method includes S101-S106, specifically as follows:

[0057] S101. Obtain sample characteristic data corresponding to a plurality of sample sealing rings respectively.

[0058] In the embodiment of the present application, each sample sealing ring corresponds to one sample characteristic data, therefore, the sample characteristic data corresponding to each sample sealing ring among the plurality of sample sealing rings is obtained to obtain a plurality of sample characteristic data.

[0059] As an optional embodiment, the sample characteristic data corresponding to a plurality of sample sealing rings are obtained through the following...

Embodiment 2

[0087] The embodiment of this application provides a classification method for sealing rings, such as figure 2 Shown is a flowchart of a classification method for sealing rings, the training method includes S201-S203, specifically as follows:

[0088] S201. Acquire a target image of a sealing ring to be detected.

[0089] In the embodiment of the present application, the process of obtaining the target image of the sealing ring to be detected is similar to the process of obtaining the sample image corresponding to the sample sealing ring in the first embodiment above, and the process of obtaining the target image of the sealing ring to be detected can refer to the first embodiment above The process is not repeated here.

[0090] S202, based on the target image, acquire target feature data that can characterize the size characteristics of the seal ring to be detected; different feature values ​​in the target feature data correspond to different outer edge position points of t...

Embodiment 3

[0095] The embodiment of the present application provides a training device for the sealing ring classification model, such as image 3 The structural block diagram of the training device of a kind of sealing ring classification model shown, this training device comprises:

[0096] The acquiring module 301 acquires sample characteristic data respectively corresponding to a plurality of sample sealing rings.

[0097] The processing module 302 is configured to, for each sample feature data, execute: input the sample feature data into a plurality of convolutional neural networks respectively, perform convolution processing on the sample feature data, and obtain corresponding to each convolutional neural network The intermediate feature matrix of the first sample; different convolutional neural networks correspond to linear convolution kernels of different sizes.

[0098] The fusion module 303 is configured to fuse the intermediate feature matrices of the first samples respective...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention provides a sealing ring classification model training method, a sealing ring classification method and device. The sealing ring classification model training method provided by the invention obtains a sealing ring classification model through training, the sealing ring classification model comprises a plurality of convolutional neural networks and classifiers, and the classifiers areconnected with the convolutional neural networks through the convolutional neural networks. Convolution processing is performed on each part of sample feature data through a plurality of convolutional neural networks, so that a first sample intermediate feature matrix corresponding to each part of sample feature data can be obtained more accurately; therefore, the second sample intermediate feature matrix obtained after fusion processing of the first sample intermediate feature matrix is accurate, and the trained sealing ring classification model is accurate based on the second sample intermediate feature matrix corresponding to each part of sample feature data. According to the sealing ring classification model obtained after training, when the sealing rings are classified, the classification accuracy is high, the types of the sealing rings can be accurately detected, and then the sealing ring detection accuracy is improved.

Description

technical field [0001] The present application relates to the technical field of computer vision, in particular to a training of a sealing ring classification model, a method and a device for classifying sealing rings. Background technique [0002] After the sealing ring is produced or before use, it is necessary to inspect the sealing ring to determine whether the sealing ring has burrs. If there are burrs, the sealing ring is not a good product; if there is no burr, it is a good product. Wherein, the sealing ring includes an O-ring. [0003] In the prior art, since the distance from the edge of the good O-ring to the center of the O-ring is a fixed value, it is possible to determine whether the O-ring has burrs by detecting the distance between the point on the edge of the O-ring and the center of the circle. If the distance from the point to the center of the O-ring exceeds the set threshold range on the edge of the O-ring, it is considered that the O-ring has a burr, th...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/045G06F18/241G06F18/253G06F18/214
Inventor 黄永祯徐栋于仕琪王凯
Owner WATRIX TECH CORP LTD
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products