一种针对光晕优化的暗通道去雾算法

A Halo-Free Image Dehazing Algorithm Based on Dark Channel Prior

为了更好解决传统暗通道去雾算法在景深突变区域容易产生光晕的问题,提出一种针对光晕优化的去雾算法。算法首先对输入图像的灰度图进行边缘检测并膨胀;其次,根据得到的边缘图将输入图像划分为边缘区域和非边缘区域;最后,自适应选择窗口大小来进行去雾处理;实验结果表明改进算法可以有效提高景深突变区域的去雾效果。

In order to tackle the problem that the halo artifacts tend to appear in edge regions with large depth jumps in the traditional dehazing al-gorithms based on dark channel prior（DCP）, proposes a halo-free single image dehazing method. Firstly, the edge image of gray-level im-age of the input foggy image can be detected by using the canny operator. Secondly, the input foggy image can be divided into edge region and non-edge region based on the edge image. Finally, the size of local window can be selected adaptively applying to dehazing process. Experimental results show that the proposed method can improve the dehazing effect of edge regions with large depth jumps effi-ciently.