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

Defective insulator sample generation method and system based on style migration method

An insulator and style technology, applied in the field of defect insulator sample generation methods and systems, can solve problems such as poor quality of defect sample generation, insufficient defect sample generation ratio, and defect samples that cannot be used for deep learning target detection model training, etc., to achieve enhanced style extraction ability, providing precision and recall, reducing the effect of change

Active Publication Date: 2021-06-01
STATE GRID SICHUAN ELECTRIC POWER CORP ELECTRIC POWER RES INST
View PDF10 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006]1. Insufficient defect sample generation ratio;
[0007]2. The quality of defective sample generation is poor;
[0008]3. Defect samples cannot be used for deep learning target detection model training

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

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0059] This embodiment provides a method for generating defective insulator samples based on the style transfer method, and the specific steps are as follows:

[0060] S1: Collect multiple insulator image samples;

[0061] S2: Divide the insulator image sample into multiple image domains according to visual differences and encode each image domain.

[0062] The specific implementation process is as follows. First, samples of defective and non-defective insulators on the power line are collected. We artificially divide them into two different image domains based on visual distinction, such as yellow style and green style image domains. Since the style belongs to the global information feature, There are only local defect differences between defective samples and non-defective samples, and the difference between the two is small in terms of style transfer tasks.

[0063] Use (0,1,...) to encode each image field, 0, 1 represent the first image field and the second image field in...

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 defective insulator sample generation method and system based on a style migration method, and the method comprises the steps: dividing a plurality of collected insulator image samples into a plurality of image domains according to visual differences, carrying out the coding of each image domain, and carrying out the style migration training of the image domains through a style migration network, obtaining a style migration device between any two image domains; and finally, performing style migration on the defective insulator sample in the image domain by using the obtained style migration device to generate a new defective insulator sample. According to the invention, the style migration is carried out on the defective insulator sample in the image domain through the style migration device, a new vivid style migration image sample is generated, the quality of the generated defective insulator image sample is high, the semantic connection information of the insulator sample is retained, and the generated defective insulator sample can effectively provide the accuracy and recall rate of the target detection model based on deep learning, and has certain practical value.

Description

technical field [0001] The present invention relates to the technical field of sample generation, in particular to a method and system for generating defect insulator samples based on a style transfer method. Background technique [0002] In recent years, in the intelligent inspection work of the power system, the State Grid has tried to introduce novel artificial intelligence technology to solve labor-intensive problems in the inspection work, such as realizing the intelligent inspection target through the target detection algorithm based on deep learning, but due to The frequency of partial fault defects in power systems is low, resulting in fewer defect samples that can be collected, and it is difficult to meet the data volume requirements of deep learning. Therefore, a method of generating defect samples is urgently needed to solve the above problems. [0003] For defective insulator samples, the amount of currently available data still does not meet the needs of neural ...

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): G06T7/00G06T3/00G06K9/62G06T7/11
CPCG06T7/0004G06T7/11G06T2207/20081G06T2207/20084G06F18/214G06T3/04
Inventor 张凌浩闫志杰唐勇梁晖辉陈亮张菊玲向思屿刘姗梅潘文分
Owner STATE GRID SICHUAN ELECTRIC POWER CORP ELECTRIC POWER RES INST
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