基于贝叶斯理论的湍流退化图像复原方法研究

余国亮; 张天序; 洪汉玉; 王宁宇

doi:10.11834/jig.200509213

学术论文与技术报告 | 浏览量 : 0 下载量: 294 CSCD: 0

PDF
导出
分享
收藏
专辑

基于贝叶斯理论的湍流退化图像复原方法研究
Investigation on Restoration Method for Turbulence-degraded Image Using Bayes Theorem
2005年10卷第9期页码：1171
纸质出版：2005
DOI： 10.11834/jig.200509213
稿件说明：

移动端阅览

余国亮, 张天序, 洪汉玉, 王宁宇. 基于贝叶斯理论的湍流退化图像复原方法研究[J]. 中国图象图形学报, 2005,10(9):1171. DOI： 10.11834/jig.200509213.

Investigation on Restoration Method for Turbulence-degraded Image Using Bayes Theorem[J]. Journal of Image and Graphics, 2005, 10(9): 1171. DOI： 10.11834/jig.200509213.

摘要

大气湍流退化图像的复原在航天成像、天文观测等领域具有重要的地位

也是目前亟需解决的问题。该问题的解决能够克服大气湍流扰动带来的图像降质和提高目标图像的分辨能力

以便于后续的目标特征提取和识别等处理。为了对湍流退化图像进行有效复原

提出了一种基于贝叶斯理论的单帧双重循环盲目去卷积图像复原算法

并对该算法的快速实现进行了研究

最后还进行了稳健性分析与测试。实验结果表明

该算法具有较强的稳定性和抗噪声能力

对于缺乏先验知识的情况尤为适用

可见该算法具有实用价值。

Abstract

Restoration of atmospheric turbulence-degraded image is very important in the field of astronomical imaging and astroobservation.It needs to be solved as soon as possible.Solving this problem can deblur the atmospheric turbulence-degraded image and improve the capability of object identification which is good for late stage works such as object feature extraction and recognition.In order to restore the turbulence-degraded image efficiently

a single frame blind deconvolution image restoration algorithm with two circulations based on Bayes theorem is presented.Its fast implementation is studied and some experiments to analyze the stability are carried out.The experimental results show that the algorithm is capable of resisting-noise with robustness;especially it is more suitable for the situation without any prior knowledge.So the algorithm possesses practical value.