基于混合暗通道算法的图像去雾研究

Image Dehazing Based on Mixed Dark Channel Prior

基于混合暗通道算法对图像去雾进行了研究。首先通过阈值划分混合暗通道,混合暗通道为近景区域和远景区域之和,最大化相似度评价函数获得混合暗通道微调系数;接着通过像素点与区域中心点的暗通道光强差值实现近景区域去雾,远景区域去雾规则是通过像素的膨胀、腐蚀实现;最后给出了算法流程。实验仿真显示该算法去雾图像视觉比其它算法的效果好,定量分析指标较优。

The image dehazing is researched based on mixed dark channel prior.Firstly,threshold was divided mixed dark channel,mixed dark channel prior was combined with near and far scene,and evaluation function of maximum similarity was obtained fine-tuning coefficient of mixed dark channel prior.Secondly,near scene was used light intensity difference of dark channel of pixel point and region center,and far scene was used expansion and corrosion of pixel.Finally,the algorithm process was given.Experimental simulation shows that the mixed dark channel prior has better visual effect and quantitative analysis indexes than other algorithms.