EN
Submit
Register
Online Submission Pre-typesetting Academic Support
Home > Wen Yuan (Middle School Edition) > 2021 No.12 > 基于生成对抗网络的焊缝X射线样本生成方法研究

  • 基于生成对抗网络的焊缝X射线样本生成方法研究

    Author(s)：

    王雪 黄旭 邹懿

    Publication Title：

    Wen Yuan (Middle School Edition) 2021 No.12

    Affiliations：

    重庆科技学院电气工程学院，重庆 401331

    Abstract：

    研究压力容器焊缝检测中训练样本的生成问题，训练样本过少会导致检测模型训练不充分，影响检测精度。对WGAN（Wasserstein-GAN）使用AdaBelief优化器进行优化，代替随机梯度下降，提升训练GAN的稳定性，提高生成样本的质量。以解决压力容器焊缝缺陷样本质量差，数量少的问题。通过GDX-ray焊缝图像生成实验证明，使用AdaBelief优化器的WGAN能够在小样本的情况下生成高质量的压力容器焊缝缺陷样本。

    Keywords：

    焊缝；小样本；生成对抗网络；AdaBelief

    Views
    6
    Citations
    0
    Downloads
    0
    Favorites
    0
    Full Text
    Download PDF
    Cite
    Favorited
    Favorite
    Share via WeChat

    Scan to share on WeChat

    Mobile Reading

    Scan to read on mobile

    Cite
    Copy

    [1] 王雪,黄旭,邹懿.基于生成对抗网络的焊缝X射线样本生成方法研究[J].Wen Yuan (Middle School Edition),2021(12),.

文渊(中学版)
Wen Yuan (Middle School Edition)

Language: 中文

CN：11-9275/G

ISSN：2096-627X

Publication Frequency: 月刊

Recommended Articles
Friendship Link

An academic processing and support platform focusing on journal databases, submission and review, and pre-typesetting.

OpenScirbe Publishing
Pre-typesetting
Article Retrieval
News
Latest news
Industry news
Academic Support
Language editing
Translation
International dissemination
Journal
Wen Yuan (Primary School Edition)
Wen Yuan (Middle School Edition)
Wen Yuan (High School Edition)
Urban and Rural Studies
Contact Us
027-59496650
Office
Room 510, Office Tower, Optics Valley Financial Center, 668 Gaoxin Avenue, Donghu New Technology Development Zone, Wuhan, Hubei Province
User Agreement Privacy Policy
鄂ICP备2026003587号-1 copyright 2026 All rights reserved. Tel:027-59496650 鄂公网安备42018502009073号