暗通道先验结合直方图均衡的阴雾图像处理方法

Image Dehazing Using Dark Channel Prior with Histogram Equalization

为解决暗通道先验对低亮度阴雾图像去雾后的图像亮度较暗,图像内的物体较模糊,不易识别的问题,提出暗通道先验结合直方图均衡的阴雾图像处理方法;利用数学软件,分别测试单独利用暗通道先验、单独利用直方图均衡、暗通道先验结合直方图均衡这3种方法处理阴雾图像的效果。结果表明,相较于单独使用暗通道先验与直方图均衡处理的阴雾图像,所提出的暗通道先验结合直方图均衡方法处理的阴雾图像平均梯度分别为前2种方法的2.18倍与1.63倍,信息熵分别为前2种方法的1.12倍与0.99倍,采用所提方法处理的阴雾图像综合指标,比单独使用暗通道先验与直方图均衡处理的阴雾图像综合指标更高。

To address the difficult recognition of low-light hazy images because of dark and vague state after dehazing,an image dehazing approach using dark channel prior combined with histogram equalization is proposed in this paper.Mathematical software is also used to test the performance of hazy images processed by dark channel prior alone,histogram equalization alone and dark channel prior combined with histogram equalization.The results show that the average gradient of the low-brightness hazy images processed by this method is 2.18 times and 1.63 times higher than that of the low-brightness hazy images processed by the dark channel prior and histogram equalization alone,its information entropy is 1.12 times and 0.99 times higher,and its composite index is higher too.