结合直方图均衡化和暗通道先验的去雾算法

摘要

Aiming at problem that some hazes cannot be removed completely from a single image based on dark channel prior and processing speed is slow,an improved algorithm is proposed combined with histogram equalization algorithm.The algorithm performs histogram equalization on the minimum image of the fog image in order to improve the contrast of the minimum image.The processed minimum image is filtered utilizing the bilateral filtering characteristics of edge-preserving smoothing,which can be used as guide map to refine the initial transmission map.The average value of the pixel values in the dark channel images at the first 0.1% is taken as the atmospheric light value A,and the defogging image is restored according to the atmospheric scattering model. The experimental results show that the transmissivity image obtained using the algorithm has better sharpness and the computational speed is greatly improved.Meanwhile,the algorithm compensate the weakness of transmissivity estimation in bright area.%针对暗通道先验单幅图像去雾算法去雾不彻底且速度慢的问题,提出了一种结合直方图均衡化算法的改进算法.在分析光晕产生的基础上,算法对有雾图像的最小值图像进行了直方图均衡化,提高了最小值图像的对比度;利用双边滤波平滑且保持边缘的特性细化对比度增强后的最小值图像,以其作为引导图对初始透射率图细化.同时,取暗通道图像中强度值在前0.1%的像素点的值求平均值,作为大气光值A,根据大气散射模型恢复无雾图像.实验结果表明:算法得到的透射率图恢复的图像具有更好的清晰度,提高了运算速度,弥补了传统算法在明亮区域透射率估计的不足.