基于暗通道先验的图像去雾算法改进

发布时间：2018-02-23 22:18

本文关键词： 去雾暗通道先验估计光线传播图 Sobel算子　出处：《液晶与显示》2016年08期 　论文类型：期刊论文

【摘要】：为了实现基于物理模型的图像复原去雾算法,文中提出了一种改进的基于暗通道先验的图像去雾算法。介绍了雾天图像退化模型和基于该雾天图像退化模型的几种去雾算法。详细介绍了何恺明提出的基于暗通道先验的去雾算法,该算法在估计光线传播图时使用的基于导向滤波的软抠图非常耗时,经过改进,直接使用景深估计光线传播图,算法运行时间大大减少。最后,使用MATLAB对改进的去雾算法进行仿真,并与原算法的运行时间进行比较。结果显示新方法对光线传播图的估计可靠,运行时间对比改进前大约下降60%,实时性大大提高。带有天空的有雾图像去雾后色斑和光晕大幅减少,取得了很好的效果。改进的去雾算法运行速度快、去雾效果好,新提出的光线传播图估计方法可靠,并且去雾过程中得到的光线传播图可以用于其他应用。
[Abstract]:In order to realize the image restoration and de-fogging algorithm based on physical model, In this paper, an improved image de-fogging algorithm based on dark channel priori is proposed. The image degradation model of fog sky and several de-fogging algorithms based on this model are introduced. He Kaiming's algorithm based on dark is introduced in detail. Channel priori de-fogging algorithm, The soft matting based on orientation filter used in the estimation of ray-propagation graph is very time-consuming. After improvement, the algorithm uses the depth of field directly to estimate the ray-propagating graph, and the running time of the algorithm is greatly reduced. Finally, The improved de-fogging algorithm is simulated by MATLAB and compared with the original algorithm. The results show that the new method is reliable in estimating the light propagation diagram. The running time is reduced by about 60% before the improvement, and the real time is greatly improved. The color spot and halo of the fog image with sky are greatly reduced, and the effect is very good. The improved de-fog algorithm runs fast and has good effect of removing fog. The new method is reliable and can be used in other applications.
【作者单位】：中国科学院长春光学精密机械与物理研究所;中国科学院大学;
【基金】：国家863高技术研究发展计划资助项目(No.2015AA7031010B)~~
【分类号】：TP391.41

【相似文献】