首页|基于两步迭代收缩法的多稀疏空间图像快速重构方法

基于两步迭代收缩法的多稀疏空间图像快速重构方法

扫码查看
原文链接
  • 国家科技期刊平台
  • NETL
  • NSTL
  • 万方数据
由于多稀疏空间图像重构时,像素范围选取过大、峰值信噪比低以及重构时间长,导致图像重构方法存在重构效果差的问题,提出基于两步迭代收缩法的多稀疏空间图像快速重构方法.明确多稀疏空间图像重构存在的问题,在明确问题后,以迭代收缩阈值算法为基础,引入迭代加权收缩算法,结合每一轮迭代结果作为初值,完成图像重构的两步迭代收缩法设计,实现多稀疏空间图像快速重构.实验结果表明:应用该方法后的重构多稀疏空间图像峰值信噪比达到 37.9 dB以上,图像重构时间仅为16.0 ms,图像结构相似性达到了0.98以上,并且重构多稀疏空间图像的效果更好,经过实验分析证实了所提方法具备可行性.
A Fast Reconstruction Method of Multi-Sparse Space Image Based on Two-Step Iterative Shrinkage Method
Image reconstruction methods have the problem of poor reconstruction effect due to the large pixel range selection,low peak signal-to-noise ratio and long reconstruction time in multi-sparse space image reconstruction.A fast reconstruction method of multi-sparse space image based on two-step iterative contraction is proposed.The problems existing in multi-sparse space image reconstruction were clarified.The iterative weighted shrinkage algorithm was introduced based on the iterative shrinkage threshold algorithm after iden-tifying the problems,and the results of each iteration were combined as initial values,the two-step iterative shrinkage method design of image reconstruction was completed to realize the rapid reconstruction of multi-sparse space image.The experimental results show that the peak signal-to-noise ratio of reconstructed multi-sparse space images can reach more than 37.9 dB,image reconstruction time is only 16.0 ms,the image structure similarity can reach more than 0.98,and the reconstruction effect of multi-sparse space images is better.The feasibility of method has been verified by experimental analysis.

multi-sparse space imageimage reconstructiontwo-step iterative shrinkage methoditerative weighted shrinkage algorithmiterative shrinkage threshold algorithm

许学添、郑禹

展开 >

广东司法警官职业学院信息管理系,广东 广州 510520

安徽建筑大学机械与电气工程学院,安徽 合肥 230601

多稀疏空间图像 图像重构 两步迭代收缩法 迭代加权收缩算法 迭代收缩阈值算法

2022年中国高校产学研创新基金新一代信息技术创新项目2023年广东省普通高校特色创新类项目国家自然科学基金青年科学基金

2022IT0722023KTSCX29581600808

2024

10.3969/j.issn.1005-9490.2024.01.025

电子器件
东南大学

电子器件

CSTPCD
影响因子：0.569
ISSN：1005-9490
年,卷(期)：2024.47(1)
  • 15