Restoration of misty images based on wavelet transform and atmospheric scattering model

IF 1.4 4区物理与天体物理 Q4 MATERIALS SCIENCE, MULTIDISCIPLINARY

AIP Advances Pub Date : 2024-08-30 DOI:10.1063/5.0219235

Pengcheng Liu

{"title":"Restoration of misty images based on wavelet transform and atmospheric scattering model","authors":"Pengcheng Liu","doi":"10.1063/5.0219235","DOIUrl":null,"url":null,"abstract":"In conditions of low visibility, the scattering effect of atmospheric particles can result in images captured by cameras exhibiting a number of problems, including reduced contrast, color distortion, and blurred details. These issues can have a significant impact on the clarity and recognition of the images. In response to these issues, this study combines atmospheric scattering models with wavelet transform, introduces non-blind deconvolution algorithms and, finally, designs a special gamma correction function to ultimately develop a new foggy image restoration model. The relevant outcomes showed that when the dehazing model processed most image samples, its edge preservation was mainly concentrated in a relatively high range, namely [0.80, 0.96]. From the perspective of the structural similarity index, the discrete wavelet transform achieved the highest value of 0.8264. In addition, the peak signal-to-noise ratio of the improved gamma correction function was 18.346. The proposed algorithm for restoring foggy images has the advantages of high clarity, strong color restoration ability, and fast restoration speed, providing more reliable and efficient image data support for applications in related fields.","PeriodicalId":7619,"journal":{"name":"AIP Advances","volume":null,"pages":null},"PeriodicalIF":1.4000,"publicationDate":"2024-08-30","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"AIP Advances","FirstCategoryId":"88","ListUrlMain":"https://doi.org/10.1063/5.0219235","RegionNum":4,"RegionCategory":"物理与天体物理","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q4","JCRName":"MATERIALS SCIENCE, MULTIDISCIPLINARY","Score":null,"Total":0}

引用次数: 0

Abstract

In conditions of low visibility, the scattering effect of atmospheric particles can result in images captured by cameras exhibiting a number of problems, including reduced contrast, color distortion, and blurred details. These issues can have a significant impact on the clarity and recognition of the images. In response to these issues, this study combines atmospheric scattering models with wavelet transform, introduces non-blind deconvolution algorithms and, finally, designs a special gamma correction function to ultimately develop a new foggy image restoration model. The relevant outcomes showed that when the dehazing model processed most image samples, its edge preservation was mainly concentrated in a relatively high range, namely [0.80, 0.96]. From the perspective of the structural similarity index, the discrete wavelet transform achieved the highest value of 0.8264. In addition, the peak signal-to-noise ratio of the improved gamma correction function was 18.346. The proposed algorithm for restoring foggy images has the advantages of high clarity, strong color restoration ability, and fast restoration speed, providing more reliable and efficient image data support for applications in related fields.

查看原文本刊更多论文

基于小波变换和大气散射模型的迷雾图像复原

在能见度低的条件下，大气颗粒的散射效应会导致照相机拍摄的图像出现一系列问题，包括对比度降低、色彩失真和细节模糊。这些问题会对图像的清晰度和识别度产生重大影响。针对这些问题，本研究将大气散射模型与小波变换相结合，引入了非盲解卷算法，并设计了一种特殊的伽玛校正函数，最终开发出一种新的雾图像修复模型。相关结果表明，当去毛刺模型处理大部分图像样本时，其边缘保留主要集中在一个相对较高的范围内，即[0.80, 0.96]。从结构相似性指数来看，离散小波变换的相似性指数最高，达到 0.8264。此外，改进的伽玛校正函数的峰值信噪比为 18.346。所提出的雾图像修复算法具有清晰度高、色彩还原能力强、修复速度快等优点，为相关领域的应用提供了更可靠、更高效的图像数据支持。

本文章由计算机程序翻译，如有差异，请以英文原文为准。

求助全文

约1分钟内获得全文求助全文

来源期刊

AIP Advances NANOSCIENCE & NANOTECHNOLOGY-MATERIALS SCIENCE, MULTIDISCIPLINARY

CiteScore

2.80

自引率

6.20%

发文量

1233

审稿时长

2-4 weeks

期刊介绍： AIP Advances is an open access journal publishing in all areas of physical sciences—applied, theoretical, and experimental. All published articles are freely available to read, download, and share. The journal prides itself on the belief that all good science is important and relevant. Our inclusive scope and publication standards make it an essential outlet for scientists in the physical sciences. AIP Advances is a community-based journal, with a fast production cycle. The quick publication process and open-access model allows us to quickly distribute new scientific concepts. Our Editors, assisted by peer review, determine whether a manuscript is technically correct and original. After publication, the readership evaluates whether a manuscript is timely, relevant, or significant.