一种有效的图像去雾方法  

An Effective Method for Image Defogging

作　　者：王建富[1,2] 钮俊清[1,2] 任清安[1,2] 

机构地区：[1]中国电子科技集团公司第三十八研究所,合肥230088 [2]孔径阵列与空间探测安徽省重点实验室,合肥230088

出　　处：《遥测遥控》2017年第2期32-35,共4页Journal of Telemetry,Tracking and Command

摘　　要：使用一种有效的方法实现雾化图像的去雾处理。首先根据雾天图像退化的物理模型,基于暗原色先验的单一图像去雾算法原理实现图像去雾的粗处理;然后针对粗处理后的图像,提出一种简单的去块滤波方法,有效地消除块效应;最后将图像转换到HSV颜色模型下,对亮度进行处理,降低色彩失真。实验结果表明,我们所提出的方法能够显著地改善有雾图像的视觉效果。In this paper, we propose a method which can effectively remove fog effect from images. Based on the model of foggy image degrading and the principle of dark channel priority, foggy images are processed roughly. Then, a simple deblocking filter is used to remove the blocking effects. Finally, under HSV color space, we adopt nonlinear transformation of brightness to reduce the distortion in color. The algorithm has been evaluated on images with different foggy concentration under different scenes. Experimental results show that the proposed approach can improve the visual effect of image well

关 键 词：图像去雾 去块滤波 暗原色 图像退化 

分 类 号：TP37[自动化与计算机技术—计算机系统结构]