Research on restoration technology of motion blur image

CHEN Ying; HONG Chenfeng

doi:10.13873/J.1000-9787(2021)04-0063-03

Chinese

您当前的位置：

首页 >

文章列表页 >

Research on restoration technology of motion blur image

更新时间：2025-09-11

- Research on restoration technology of motion blur image
- Issue 4, Pages: 63-65+73(2021)
- 作者机构：
  
  南昌航空大学软件学院
- 作者简介：
- 基金信息：
- DOI：10.13873/J.1000-9787(2021)04-0063-03
  CLC： TP391.41
- Published：2021
- 稿件说明：
移动端阅览
CHEN Ying, HONG Chenfeng. Research on restoration technology of motion blur image[J]. 2021, (4): 63-65+73.
DOI：

CHEN Ying, HONG Chenfeng. Research on restoration technology of motion blur image[J]. 2021, (4): 63-65+73. DOI： 10.13873/J.1000-9787(2021)04-0063-03.

摘要

针对运动模糊图像的恢复

提出了基于生成式对抗神经网络(GAN)网络与FSRCNN网络的方案。采用Go Pro模糊图像数据集与DIV2K图像恢复数据集进行网络训练;通过图像处理对模糊图像进行规格化、归一化、色彩空间转换等预处理操作;利用GAN对模糊图像进行图像恢复

并结合FSRCNN针对模糊图像进行图像增强;针对GAN与FSRCNN的处理结果进行分析与对比

经由FSRCNN网络图像增强的恢复图像的峰值信噪比虽然有一定的下降

但结构相似度则得到了提升。实验结果表明本文的算法方案具有较好的可行性。

Abstract

An image restoration scheme based on generative adversarial networks（ GAN） network and FSRCNN network for motion blurred images is proposed.Go Pro fuzzy image dataset and the DIV2 K image restoration dataset are used for network training.Image processing operation is used to normalize the fuzzy image

convert color space.GAN is used to restore the image of the blurred image

and the FSRCNN is combined to enhance the image for the blurred image.Processing results of GAN and FSRCNN are analyzed and compared.The peak signal-noise ratio of the reconstructed image enhanced by the FSRCNN network image has a certain decline.However

the structural similarity is improved.Experimental results show that the proposed algorithm has good feasibility.

关键词

Keywords

references

Views

408

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Road surface crack detection based on CNN

Lightweight bubble detection algorithm for quartz crucibles based on improved YOLOv5

Research on improved median filtering algorithm fused with gradient information

Indoor localization algorithm based on image processing and RSSI signal

Traffic sign recognition algorithm based on double-channel CNN

Related Author

ZHAO Shan-shan

HE Ning

ZHAO Qian

ZHENG Xuan

ZHAO Man

MA Xiang

DU Zhonghua

CAI Yu

Related Institution

Beijing Key Laboratory of Information Services Engineering,Beijing Union University

College of Intellectualized City,Beijing Union University

School of Communication and Information Engineering,Xi’an University of Science and Technology

2.Xi’an Dishan Vision Technology Limited Company

School of Mechanical Engineering,Nanjing University of Science and Technology

⁰