摘要
提出了一种适用于恶劣天气下退化图像复原的简化算法。该算法利用大气辐射传输方程的数值解构造点扩散函数,可用于不同天气条件下的退化图像的复原。根据恶劣天气中散射粒子的前向散射系数大的特点,引入δ 函数,将相函数进行归一化分解;把得到的简化点扩散函数作为大气的退化函数,与退化图像一起进行频域复原滤波。实验结果表明,对于 512×512 的雾天场景图像,复原处理前后信噪比从 3.6400 提高到 8.5329;比较具有相同信噪比的复原图像,简化算法的处理耗时比简化前减少了 85%。
A simplified algorithm suitable for restoring fully degraded images under bad weather conditions is proposed. The numerical value solution of atmospheric radiation transmission equation (RTE) is used to construct point spread function and implement the restoration of all kinds of atmosphere degraded images under different weather conditions. The function δ is introduced according to the property of the large forward scattering coefficient of the scattering particles in bad weather and the decomposition for phase function is normalized. By using the obtained simplified point spread function as atmosphere degraded function and together with the degraded image, one can restore the image through filtering in frequency domain. The experimental results show that for a 512×512 scene image in foggy day, the signal-noise-ratio gained by the new algorithm has been improved from 3.6400 to 8.5329 in comparison with the image before restoration and the processing time consumption decreases by 85% than before.
出处
《光电工程》
EI
CAS
CSCD
北大核心
2005年第1期71-73,共3页
Opto-Electronic Engineering
关键词
图像复原
信噪比
相函数
多重散射
Image restoration
Signal-noise ratio
Phase function
Multiple scattering
