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

A two-stage mobilenet-based defect detection method for bathroom ceramics

A technology for defect detection and ceramics, applied in neural learning methods, image analysis, image enhancement, etc., can solve the problems of detection speed requirements, detection accuracy requirements, etc., to meet the requirements of detection speed, good detection effect, and high detection speed Effect

Active Publication Date: 2022-07-05
HARBIN INST OF TECH
View PDF5 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the present invention is to solve the problem that the existing sanitary ceramics defect detection method cannot meet the detection speed requirement while meeting the detection accuracy requirement, and proposes a sanitary ceramics defect detection method based on two-stage MobileNet

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 two-stage mobilenet-based defect detection method for bathroom ceramics
  • A two-stage mobilenet-based defect detection method for bathroom ceramics
  • A two-stage mobilenet-based defect detection method for bathroom ceramics

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0018] Embodiment 1: A method for detecting defects of bathroom ceramics based on two-stage MobileNet in this embodiment includes the following steps:

[0019] Step 1: Build and train a defect localization network to obtain a trained defect localization network, including the following steps:

[0020] Step 11. Build a defect location network (such as figure 1 ):

[0021] The defect localization network includes: a front-stage feature extraction network and a three-layer convolutional network of the defect localization network;

[0022] The front-level feature extraction network of the defect localization network adopts the backbone network of MobileNetV3;

[0023] The backbone network of MobileNetV3 includes: a pilot convolution operation (3*16 convolution, data normalization, hswish activation), a bneck structure (including fifteen block components, each block component is composed of three layers of spinning cones. Shaped convolution operation module) and a tail convoluti...

Embodiment

[0058] In this embodiment, the defect samples collected at the production site of sanitary ceramics are used to test the effectiveness of the method. The defect types and the number of samples are shown in Table 1, and the size of the sample is 9.6 million pixels. During the testing process, 60% of the original data is used to construct the training set, and the remaining 40% of the original data is used as the test set. The calculation formula of the detection accuracy is as follows:

[0059]

[0060] The test results are shown in Table 2. The average detection frame rate reached 18FPS.

[0061] Table 1 Defect data statistics table

[0062]

[0063] Table 2 Defect detection accuracy statistics table

[0064]

[0065] It can be seen that for the defects of sanitary ceramics, the present invention not only has a high detection accuracy, but also has a detection speed that can meet the needs of the production site.

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

A two-stage MobileNet-based bathroom ceramic defect detection method relates to the fields of industrial automation, artificial intelligence and machine vision. The present invention is to solve the problem that the existing bathroom ceramic defect detection method cannot satisfy the detection accuracy requirement and the detection speed requirement at the same time. The invention includes: inputting a bathroom ceramic image to be detected into a trained defect location network to obtain a heat map of the bathroom ceramic to be detected, and extracting the defect location result to be detected by using the heat map of the bathroom ceramic to be detected; the defect location The network includes: a pre-feature extraction network and a multi-layer convolution network of the defect localization network; the pre-feature extraction network of the defect localization network adopts the backbone network of MobileNetV3; the obtained defect localization results to be detected are input into the trained The defect detection results are obtained in the defect classification network. The invention is used for detecting whether sanitary ceramics are defective.

Description

technical field [0001] The invention relates to the fields of industrial automation, artificial intelligence and machine vision, in particular to a two-stage MobileNet-based bathroom ceramic defect detection method. Background technique [0002] In modern society, the use of ceramic tiles is a common decoration method to decorate buildings. High-grade house buildings often use more advanced ceramic tiles, while bathrooms use bathroom tiles that are different from other ceramic tiles. However, with the rise of the ceramic tile industry, people's requirements for the quality of ceramic tiles are also getting higher and higher. In the face of competition in the industry, the ceramic tile industry is also under more pressure. Product quality inspection is an indispensable link in every production line. With the development of artificial intelligence and machine vision technology, automatic inspection technology based on machine vision forms and artificial intelligence methods is...

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): G06T7/00G06T7/73G06V10/764G06V10/774G06V10/82G06K9/62G06N3/04G06N3/08
CPCG06T7/0008G06T7/73G06N3/08G06T2207/20081G06T2207/20084G06N3/045G06F18/2431G06F18/214
Inventor 高会军杭景帆杨宪强
Owner HARBIN INST OF 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