Unlock instant, AI-driven research and patent intelligence for your innovation.

A Dataset Augmentation Method for Visual Detection of Appearance Defects

A visual inspection and appearance defect technology, applied in the field of visual inspection, can solve the problems of model overfitting, low degree of diversification, insufficient number of samples, etc., and achieve the effect of high-precision surface defect detection

Active Publication Date: 2021-03-12
苏州佳赛特智能科技有限公司
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] To apply machine learning to actual industrial inspection, it is necessary to solve the problems of insufficient number of defect samples and low degree of diversity in the training stage of machine learning
In the case of insufficient samples, using Generative Adversarial Networks (GAN) to amplify the data set can effectively solve the problems of model overfitting and low detection accuracy caused by insufficient training samples, but the conventional GAN It is difficult for the model to generate high-quality training samples

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
  • A Dataset Augmentation Method for Visual Detection of Appearance Defects
  • A Dataset Augmentation Method for Visual Detection of Appearance Defects
  • A Dataset Augmentation Method for Visual Detection of Appearance Defects

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0027] The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments, so that those skilled in the art can better understand the present invention and implement it, but the examples given are not intended to limit the present invention.

[0028] refer to Figure 1-2 As shown, a data set amplification method for visual detection of appearance defects, the specific steps include:

[0029] S1. Acquiring images for visual detection, and performing block processing on the acquired images as a training data set;

[0030] S2. The generator in the classic Generative Adversarial Network (GAN) adopts a deconvolutional neural network, integrates the image defect enhancement module into the generation confrontation network, and adds a feedback channel at the front end of the output bias and image defect enhancement module;

[0031] S3, the defect sample obtained after the training data set is processed is used as a training s...

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 relates to the technical field of visual inspection, and relates to a data set amplification method for visual inspection of appearance defects. The present invention can generate high-definition and high-diversity defect data, which uses generative confrontation network to amplify high-quality data sets through very limited defect samples, so that the amplified data sets can support machine learning training , to achieve high-precision surface defect detection using machine learning even in the case of limited defect samples.

Description

technical field [0001] The invention relates to the technical field of visual inspection, and relates to a data set amplification method for visual inspection of appearance defects. Background technique [0002] With the vigorous development of the economy, my country's manufacturing industry is also developing rapidly, and higher and higher requirements are put forward for the appearance quality of industrial products. The traditional surface defect detection method is manual visual inspection. At present, most manufacturers still use manual visual inspection to identify defective products, which is inefficient and often leads to missed inspections. Resulting in lower product quality, resulting in a waste of resources. In order to solve the problems caused by manual visual inspection, intelligent detection methods based on machine vision will gradually replace manual visual inspection in the production line quality appraisal process and become the mainstream method. [00...

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 Patents(China)
IPC IPC(8): G06T7/00
CPCG06T7/0004G06T2207/20081G06T2207/20084
Inventor 何志勇林嵩
Owner 苏州佳赛特智能科技有限公司