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

Unsupervised enhancement method for surface defect image data of hot-rolled plate coil

An image data and hot-rolled coil technology, which is applied in the field of image data generation of hot-rolled coil surface defect image data based on artificial intelligence network model, can solve the problem of affecting product quality, less picture data, and low accuracy of defect recognition by artificial intelligence machine vision technology. And other issues

Pending Publication Date: 2020-11-24
AUTOMATION RES & DESIGN INST OF METALLURGICAL IND
View PDF0 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a method that can stably and effectively generate image data of surface defects of hot-rolled coils, so as to solve the problem that the data of surface defects of hot-rolled coils in the field of metallurgical rolling are scarce and difficult to obtain, which leads to the use of artificial intelligence machine vision technology to identify The accuracy of defects is low, which seriously affects the quality of its products

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
  • Unsupervised enhancement method for surface defect image data of hot-rolled plate coil
  • Unsupervised enhancement method for surface defect image data of hot-rolled plate coil
  • Unsupervised enhancement method for surface defect image data of hot-rolled plate coil

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0043] 1. The implementation of this method requires a computer server for deep learning, 2 GPU graphics cards of Tesla (version V100), each with a memory size of 32G, 128 internal memory, and 1TB solid-state hard drive as the hardware platform;

[0044] 2. Install the Linux 64-bit (Ubuntu16.04) operating system on the computer, install the Anaconda software library, install the Pytorch deep learning framework, install the CUDAToolkit version 10.2, the graphics driver version is NVIDI-440.33.01, and the cudnn acceleration package version is 7.6.5 As a software environment, use python programming language;

[0045] 3. Use the command statement source axtivatepytorch to activate the development environment for deep learning;

[0046] 4. Prepare the data set. The picture training set in the present invention comes from the actual production line of hot-rolled coils in the steel mill, and is photographed and collected by the surface defect detector. The labeled data set after work...

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 discloses an unsupervised enhancement method for hot-rolled plate coil surface defect image data, and belongs to the technical field of metallurgical hot-rolled plate coil defect detection. The generative adversarial network is composed of two deep neural networks with mutual adversarial competition and is a probability generation model, forward propagation is carried out on a samplegeneration mode through a generator, gradient back propagation is used for carrying out optimization calculation after discrimination is carried out through a discriminator, and the method does not depend on any prior hypothesis. In the training process of the productive adversarial network, the two networks are continuously iterated and optimized in the mutual game, the generator learns and generates a more real sample, and the discrimination network discriminates whether a data sample is a real sample or a generated false sample as much as possible. The two parties continuously compete until the last two parties cannot become better, and finally, the two networks achieve a dynamic balance, i.e., the distribution of the images generated by the generator is close to the distribution of real images, and the discriminator cannot identify true and false images. The method has the advantage of solving the problems of few hot-rolled plate coil surface defect picture data, difficulty in collection, low recognition accuracy and the like in the metallurgical steel rolling field.

Description

technical field [0001] The invention belongs to the technical field of defect detection of metallurgical hot-rolled coils, and relates to a method for generating image data of surface defects of hot-rolled coils based on an artificial intelligence network model. Background technique [0002] In the era of continuous industrial development, steel can be said to be the "food" of industry, and it is one of the most important materials in modern times. However, in the process of actually producing steel plates and coils in steel mills, due to the influence of factors such as equipment problems and processing techniques, hot-rolled coils with surface defects are often produced. If the surface defects of the coil cannot be detected and identified effectively, it will seriously affect its quality as a steel product, its appearance, economic benefits, performance, etc., so the research on the defect identification of the coil surface has great practical significance. Early defect d...

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
IPC IPC(8): G06T7/00G06K9/62G06N3/04G06N3/08
CPCG06T7/0004G06N3/08G06T2207/20081G06N3/045G06F18/214
Inventor 杨永刚张云贵邓泽先
Owner AUTOMATION RES & DESIGN INST OF METALLURGICAL IND
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