基于三维块匹配的红外图像降噪与缺陷量化方法
DOI:
作者:
作者单位:

作者简介:

通讯作者:

中图分类号:

基金项目:


Infrared image denoising and defect quantification based on 3D block matching
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    对金属表面的缺陷检测以及量化逐渐成为一项重要的工作,对于保障器件安全有着重大意义。本文以钛合金为例,使用了一种以激光器为热源的的红外热成像缺陷量化方法。利用激光器可以点加热的优点对缺陷周围进行加热并对金属背面进行图像采集,为了使红外图像清晰表示并保留边缘,对红外图像进行同态滤波,变换后的红外图像进行噪声估计并使用基于三维块匹配的去噪算法将图像进行去噪,并且对去噪图像进行边缘检测以及缺陷量化。实验表明该系统可以有效的对金属材料进行缺陷检测,该红外去噪算法在峰峰值信噪比(PSNR)方面得到了显著的提升,在去除乘性噪声的同时也使图像边缘得到了有效地保留,该缺陷量化方法可以将误差控制在15 %以内。

    Abstract:

    The detection and quantification of metal surface defects has gradually become an important work,which is of great significance to the safety of devices.In this paper,titanium alloy is taken as an example,and an infrared thermal imaging defect quantification method using a laser as a heat source is used.The laser has the advantage of point heating to heat the defect and collect the image of the metal back.According to the characteristics of the noise distribution of infrared image,homomorphic filtering is applied to the infrared image,noise estimation is applied to the transformed infrared image,homomorphic filtering is applied to the infrared image,noise estimation is applied to the transformed infrared image,and the denoising algorithm based on three-dimensional block matching is used to denoise the image,and edge detection and defect quantification are applied to the denoised image.Experiments show that the system can effectively detect defects in metal materials.The infrared denoising algorithm has significantly improved the peak to peak signal-to-noise ratio (PSNR),and effectively preserved the edge of the image while removing multiplicative noise.The error of the defect quantification method can be controlled within 15 %.

    参考文献
    相似文献
    引证文献
引用本文

刘佳琪,张志杰,董宁琛,尹武良,赵晨阳.基于三维块匹配的红外图像降噪与缺陷量化方法[J].激光与红外,2020,50(10):1269~1275
LIU Jia-qi,ZHANG Zhi-jie,DONG Ning-chen,YIN Wu-liang,ZHAO Chen-yang.Infrared image denoising and defect quantification based on 3D block matching[J].LASER & INFRARED,2020,50(10):1269~1275

复制
分享
文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:
  • 最后修改日期:
  • 录用日期:
  • 在线发布日期: 2020-10-28
  • 出版日期: