基于大气MTF的退化图像复原  

Restoration of Blurred Image Based on Atmospheric MTF

作　　者：武鹏飞[1,2] 方帅[1,3] 徐青山[1] 饶瑞中[1] 

机构地区：[1]中国科学院安徽光学精密机械研究所中国科学院大气成分与光学重点实验室,安徽合肥230031 [2]中国科学院研究生院,北京100049 [3]合肥工业大学计算机与信息学院,安徽合肥230039

出　　处：《大气与环境光学学报》2011年第3期196-202,共7页Journal of Atmospheric and Environmental Optics

基　　金：国家自然科学基金(40775028)资助

摘　　要：由于大气中气溶胶和湍流的存在,致使图像质量退化,依据退化的物理过程,建立了基于大气调制传递函数(modulation transfer function,MTF)的退化模型,从而进行图像复原。首先,利用图像拍摄时刻湍流和气溶胶的相关大气参数,估算出粗略的大气MTF;其次,采用一种指数调节MTF曲线的方法,指数因子通过迭代递归的方法确定,并以复原图像平均梯度最大为截止条件,从而得到更加精准的MTF曲线;最后,给出了实验结果,并对其进行定量评价,结果表明复原后的图像在视觉效果和定量指标上都有很大提高。Due to the effect of turbulence and aerosol in atmosphere, image quality is degraded. According to physical process of degrading, the model of image degrading was built based on atmospheric modulation transfer function （MTF） and was used to restore blurred image. Firstly, Approximate atmospheric MTF was estimated using meteorological data when imaging, including turbulence and aerosol. Then an exponent adjusting method of MTF was used to get more accuracy MTF curve. The exponent adjusting factor was determined by iterative method, and the iteration was stopped when average gradient of restored image was the largest. At last, experimental result was given and assessed quantitatively, which showed that restored image was improved obviously in visual effect and quantitative specification.

关 键 词：图像复原 调制传递函数 湍流 气溶胶 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]