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

Multilevel model cloth defect detection method and system

A detection method and multi-level technology, applied in the field of pattern recognition, can solve the problems of low timeliness, low efficiency of manual inspection of defects, low accuracy, etc., to achieve a large amount of calculation, improve accuracy and real-time performance, and ensure accuracy Effect

Active Publication Date: 2017-08-01
HUAZHONG UNIV OF SCI & TECH
View PDF5 Cites 18 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] Aiming at the above defects or improvement needs of the prior art, the present invention provides a multi-level model cloth defect detection method and system. Problems such as low accuracy and low timeliness

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
  • Multilevel model cloth defect detection method and system
  • Multilevel model cloth defect detection method and system
  • Multilevel model cloth defect detection method and system

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach

[0047] The specific implementation method is: select 3000 sample images with a size of 227×227 as the learning set L1, including 1000 normal cloth images and 2000 defect images, including 500 holes, foreign objects, oil stains and crease defect images each. L1 is divided into training set L1 according to 4:1 1 and validation set L1 2 ; Compute L1 1 The 4 eigenvalues ​​of the GLCM in the 4 directions of the image, the 4 eigenvalues ​​are energy, entropy, contrast and inverse moment, and the obtained eigenvector V 11 Input the support vector machine (SVM) binary classification model for training, after a certain number of iterations, calculate L1 2 The 4 eigenvalues ​​of the GLCM in the 4 directions of the image, the 4 eigenvalues ​​are energy, entropy, contrast and inverse moment, and the obtained eigenvector V 12 Input the support vector machine (SVM) binary classification model for testing, if the test accuracy rate does not meet the requirements, continue training, and fi...

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 a multilevel model cloth defect detection method and system. The method is achieved through the steps of establishing a presort model Model 1, a convolutional neural network model FabricNet and a final classification model Model 2; collecting cloth images, segmenting the cloth images to obtain ROI images, calculating out GLCM feature values of the ROI images to form a feature vector V1; inputting the V1 into the Model 1 to judge whether the ROI images are defect images, inputting the defect images into the FabricNet if the ROI images are the defect images, and obtaining a texture feature vector V2; obtaining judgment results of the ROI images if the ROI images are not defect images; inputting the V2 into the Model2 to obtain defect classification results of the defect images; finally obtaining defect detection results of the cloth images. As the method disclosed by the invention utilizes the three models of the Model 1, the FabricNet and the Model 2, accuracy and real timeliness of cloth defect detection are improved, and the requirement of industrial production is met.

Description

technical field [0001] The invention belongs to the technical field of pattern recognition, and in particular relates to a method and system for detecting cloth defects in a multi-level model. Background technique [0002] In the production of textile industry, the detection of textiles is an important process, which determines the quality of products. Therefore, defect detection is a link that must not be ignored in the production process. However, the traditional detection of cloth defects is still evaluated by artificial vision. The long-term artificial inspection will not only affect the objective evaluation of product quality, but also be limited by the proficiency of the inspectors. With the increase of cloth production speed and the improvement of product quality requirements, manual inspection has great disadvantages, including slow detection speed, low detection rate and poor stability, which makes manual inspection less and less suitable for industrial production. ...

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): G01N21/88G06K9/32G06K9/62G06N3/08
CPCG06N3/08G01N21/8851G01N2021/8887G01N2021/8883G06V10/25G06F18/2411G06F18/24147G06F18/24155
Inventor 孙志刚万东肖力王卓
Owner HUAZHONG UNIV OF SCI & TECH
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

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com