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

CGAN-based inspection image data small sample expansion method

A technology of image data and image data sets, applied in neural learning methods, instruments, biological neural network models, etc., can solve problems such as low sample quality and inability to meet the accuracy of downstream model training

Active Publication Date: 2021-03-05
GUANGDONG POWER GRID CORP ZHAOQING POWER SUPPLY BUREAU
View PDF5 Cites 6 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] Chinese patent CN109325532A, with a publication date of 2019.02.12, discloses an image processing method for expanding data sets with small samples. Although data expansion is achieved, the sample quality of the expanded data based on this method is low, which cannot meet the requirements of downstream model training. the accuracy of

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
  • CGAN-based inspection image data small sample expansion method
  • CGAN-based inspection image data small sample expansion method
  • CGAN-based inspection image data small sample expansion method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0070] The accompanying drawings are for illustrative purposes only, and should not be construed as limiting the present invention; in order to better illustrate this embodiment, certain components in the accompanying drawings will be omitted, enlarged or reduced, and do not represent the size of the actual product; for those skilled in the art It is understandable that some well-known structures and descriptions thereof may be omitted in the drawings. The positional relationship described in the drawings is for illustrative purposes only, and should not be construed as limiting the present invention.

[0071] Such as figure 1 As shown, a small sample expansion method of inspection image data based on CGAN includes the following steps:

[0072] Step 1. Collect inspection images, judge the defect type corresponding to the inspection images based on experience, and manually mark the images to obtain the inspection image dataset; use anomaly detection algorithm to eliminate abno...

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 a CGAN-based inspection image data small sample expansion method. The method comprises the steps: firstly, collecting inspection images, judging defect types corresponding tothe inspection images, carrying out marking, and removing abnormal images in an inspection image data set by using an abnormality detection algorithm; preprocessing the cleaned inspection image data set by using a traditional image processing algorithm, and training a conditional generative adversarial network based on a convolutional neural network by using the inspection image data set to obtaina CGAN model capable of generating inspection image data of a given defect type; sampling and generating a large amount of inspection image data by using a trained CGAN model generator; and accordingto the image truth output by a discriminator, screening the generated images of which the truth is greater than a given truth threshold, and adding the screened images into the inspection image dataset to obtain an expanded inspection image data set.

Description

technical field [0001] The invention belongs to the technical field of inspection image processing of power transmission equipment, and more specifically relates to a CGAN-based small sample expansion method of inspection image data. Background technique [0002] The network structure of the distribution network is complex, the types of equipment are diverse, the distribution points are many and wide, the local environment is complex, and the security environment is relatively poor. Distribution network lines are becoming increasingly dense and inspections are becoming more difficult, but the timeliness and accuracy requirements for power inspections are constantly increasing. The current manual inspection method is relatively single, with low efficiency, high safety risks, and low defect discovery rate. It is necessary to use drone inspection methods to improve the current inefficient inspection methods, and at the same time combine deep learning and other technologies to ...

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
CPCG06N3/08G06N3/045G06F18/2411G06F18/214
Inventor 张谨立游林辉葛阳庾凌云胡峰孙仝陈政宋海龙黄达文王伟光梁铭聪黄志就谭子毅陈景尚李志鹏冯海林
Owner GUANGDONG POWER GRID CORP ZHAOQING POWER SUPPLY BUREAU
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