基于亚像素偏移的高质量鬼成像目标重构方法
DOI:
CSTR:
作者:
作者单位:

(中北大学 信息与通信工程学院,山西 太原 030051)

作者简介:

杨风暴(1968-),男,博士,教授,博士生导师,主要从事 信息融合、计算成像方面的研究.

通讯作者:

中图分类号:

基金项目:

山西省应用基础研究计划青年科技研究基金(201901D211233)和山西省研究生创新项目(2020SY368)资助项目 (中北大学 信息与通信工程学院,山西 太原 030051)


High quality ghost imaging target reconstruction based on subpixel-shifting
Author:
Affiliation:

(School of Information and Communication Engineering,North University of China,Taiyuan 030051,China)

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:

    针对欠采样次数下传统计算鬼成像重构图像分辨率 差和对比度低的问题,提出了一种基于亚像素偏移的高质量鬼成像目标重构方法。本文通过 分析不同颗粒大小散斑图的重构图像特性以及亚像素散斑偏移的特性后,采用2pixel×2p ixel的大尺寸散斑颗粒在空间位置分别进行水平、垂直、斜向偏移,结合未偏移原始散斑图 分别对未知目标进行图像重构,通过对四幅重构图像进行对应像素点灰度取平均值的方式把 各重构图像上由于亚像素偏移而存在的互补信息整合到同一幅图像中,进而实现1pixel×1pixel分辨率的成像。数值仿真及实验结果表明,以信噪比、对比度和结构相似性作为评价 指标,在相同的采样次数下,与未偏移散斑图重构图像相比,既获得了高对比度和分辨率的 重构图像,又能很大程度避免成像过程中细节信息遗漏的问题,其中信噪比、对比度和结构 相似性分别平均提高61.2%/46.3%/59.8%。

    Abstract:

    Aiming at the problem of poor resoluti on and low visibility of traditional ghost imaging reconstruction images under u nder-sampling numbers,a high-quality ghost imaging target reconstruction method based on sub-pixel shifting is proposed.In this paper,after analyzing the recon structed image characteristics of speckle patterns of different pattern sizes a nd the characteristics of subpixel speckle-shifting,the 2pixel×2pixel large-s ize speckle pattern are used to offset horizontally,vertically,and diagonally in the spatial position respectively.The unshifted original speckle pattern is u sed to reconstruct the unknown target separately,and the complementary informati on existing in each reconstructed image due to sub-pixel displacement is merged into the same image by performing the gray average of the corresponding pixels on the four reconstructed images.In the image,the imaging with 1pixel×1pixel re solution is realized.Numerical simulation and experimental results show that usi ng signal-to-noise ratio,visibility,and structural similarity as evaluation in dicators,under the same sampling numbers,compared with the unshifted speckle patte rn reconstructed image,both high visibility and resolution are obtained.Reconstr ucting the image can largely avoid the problem of missing details in the imaging process,in which the signal-to- noise ratio,visibility,and structural similari ty are improved by an average of 61.2%/46.3%/59.8% respectively.

    参考文献
    相似文献
    引证文献
引用本文

张思卿,杨风暴,王肖霞.基于亚像素偏移的高质量鬼成像目标重构方法[J].光电子激光,2021,32(3):241~250

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:2020-10-23
  • 最后修改日期:
  • 录用日期:
  • 在线发布日期: 2021-03-11
  • 出版日期:
文章二维码