基于多帧湍流退化图像的近视解卷积复原  

Myopic Deconvolution Restoration Based on Multiframe Turbulence Degraded Images

作　　者：冯晓星[1] 时东锋[1] 乔春红[1] 张鹏飞[1] 范承玉[1] 王英俭[1] 

机构地区：[1]中国科学院安徽光学精密机械研究所大气成分与光学重点实验室,安徽合肥230031

出　　处：《中国激光》2014年第9期239-244,共6页Chinese Journal of Lasers

基　　金：国家973计划(Y23D2m5531);中国科学院合肥物质科学研究院计算中心

摘　　要：根据贝叶斯和Parseval定理,引入了频域内多帧湍流退化图像的近视解卷积复原算法。以大气湍流长曝光光学传递函数作为估计的光学传递函数。根据频域代价函数的特点,提出分步牛顿法求解代价函数。本算法能够处理未匹配的多帧图像,并能获得理想的复原图像。计算机仿真多帧湍流退化图像的复原结果表明:即使多帧图像未匹配,在不同湍流强度和不同噪声情况下算法仍能复原出好的图像效果。According to the Bayes and Parseval theorem,myopic deconvolution algorithm for multiframe turbulence degraded images in frequency domain is presented.Atmosphere turbulence long exposure optical transfer function is used as the estimated optical transfer function.According to characteristics of the cost function in frequency domain,partial Newton algorithm is introduced.This algorithm can handle unregistration of multiframe images and obtain ideal recovered image.The restored result of the computer simulation multiframe turbulence degraded images shows that it has good restoration effect at different turbulence intensities and signal noises,even though the images are unregistration.

关 键 词：图像处理 大气湍流效应 光学成像 近视解卷积 图像匹配 

分 类 号：TN012[电子电信—物理电子学]