河北水利电力学院学报

河北水利电力学院学报 ›› 2023, Vol. 33 ›› Issue (3): 1-8.DOI: 10.16046/j.cnki.issn2096-5680.2023.03.001

• 人工智能与机器人专题 •    下一篇

一种电子换向器表面缺陷图像去噪和增强算法

彭宁1, 吴浩1, 漆梓渊1, 刘彦希1, 宋弘2   

  1. 1.四川轻化工大学自动化与信息工程学院,四川省自贡市汇兴路519号 643000;
    2.阿坝师范学院电子信息与自动化学院,四川省阿坝市佛山大道 624000
  • 收稿日期:2023-05-08 修回日期:2023-05-26 出版日期:2023-09-30 发布日期:2023-10-26
  • 通讯作者: 宋 弘(1973-),男,四川眉山人,教授,主要研究方向:智能信息处理。E-mail:3452974910@qq.com
  • 作者简介:彭 宁(1998-),男,安徽合肥人,硕士研究生,主要研究方向:图像处理、缺陷检测。E-mail:1797344335@qq.com
  • 基金资助:
    四川省科技厅项目(2020YFG0178,2021YFG0313,2022YFS0518,2022ZHCG0035);人工智能四川省重点实验室项目(2019RYY01);企业信息化与物联网测控技术四川省高校重点实验室项目(2019WZY02,2020WZY02);自贡市科技局项目(2019YYJC13,2019YYJC02,2020YGJC16)

An Algorithm for Image Denoising and Enhancement of Surface Defects of Electronic Commutator

PENG Ning1, WU Hao1, QI Zi-yuan1, LIU Yan-xi1, SONG Hong2   

  1. 1. School of Automation and Information Engineering, Sichuan University of Science &Engineering, 643000, Zigong, Sichuan, China;
    2. School of Electronic Information and Automation, ABA Teachers University, 624000, Aba, Sichuan, China
  • Received:2023-05-08 Revised:2023-05-26 Online:2023-09-30 Published:2023-10-26

PDF

摘要： 针对电子换向器表面缺陷图像对比度低、细节不明显、噪声过多等问题,提出一种新的电子换向器缺陷图像增强方法:针对电子换向器表面存在的混合噪声问题,通过联合中值滤波和小波阈值进行去噪;然后通过DeBlurGAN V2网络将去噪后的图像进行去模糊处理;最后以优化的非锐化掩膜(USM)和自适应对比度增强算法(ACE)构建基于非锐化掩膜的局部自适应对比度增强算法(UACE)对处理后图像中的缺陷进行对比度增强。实验结果表明,处理后图像的峰值信噪比、对比度以及信息熵均有提升,不仅消除了图像中的噪声,还有效地提升了图像的对比度。

关键词: 电子换向器, 中值滤波去噪, 小波阈值去噪, 自适应对比度增强, 非锐化掩膜

Abstract: Aiming at the problems of low contrast, obscure details and excessive noise of the surface defect image of electronic commutator in industry, a new image enhancement method for the surface defect image of electronic commutator is proposed in this paper. Firstly, aiming at the mixed noise problem on the surface of the electronic commutator, the median filter and wavelet threshold are combined to remove the noise. Then, the de-noised image is deblurred through DeBlurGAN V2 network. Finally, UACE, an adaptive contrast enhancement algorithm based on unsharp mask, is constructed with optimized USM and ACE to enhance the contrast of defects in the processed image. The experimental results show that the peak signal-to-noise ratio, contrast and information entropy of the image enhanced by the algorithm in this paper are improved, which proves that the algorithm in this paper can not only eliminate the noise in the image, but also effectively improve the contrast of the image.

Key words: electronic commutator, median filter denoising, wavelet threshold denoising, ACE, USM

中图分类号: