基于暗通道的图像去雾算法研究

Research on image haze removal algorithm based on dark channel

针对图像去雾的问题,提出了一种在Matlab平台下,基于暗通道先验的图像去雾算法。基于暗通道图像去雾的原理得到去雾图像的透射图。为了提高去雾图像的图像质量,对得到的透射图进行中值滤波处理,以削弱其方块效应。通过导向滤波对处理后的透射图进行精细化处理。为了减少滤波带来的能量损失,通过Gamma校正提高去雾后的图像亮度,从而获得最终的图像去雾结果。讨论了该算法在实际图像去雾中的可行性及去雾效果。经过多轮实验可知,该算法与经典暗通道去雾算法相比较,其信息熵(H)提高了0.12,峰值信噪比(PSNR)提高了0.7,可见边增加率(E)提高了4.8%。实验结果表明,所提出的图像去雾算法是有效的。

An image haze removal algorithm based on Dark Channel prior was proposed under the Matlab platform to solve the problem of image haze removal.The transmission map of the haze image is obtained based on the principle of dark channel image haze removal.In order to improve the image quality of the dehazing image,median filtering is applied to the obtained transmission map to weaken its block effect.Guided filtering is used to refine the processed transmission map.In order to reduce the energy loss caused by filtering,Gamma correction is used to increase the brightness of dehazing image,so as to obtain the final result.The feasibility and dehazing effect of the algorithm in practical image dehazing are discussed.After multiple rounds of experiments,the algorithm compared to the classical dark channel prior image dehazing algorithm,its information entropy(H)increased by 0.12,Peak Signal-to-Noise Ratio(PSNR)increased by 0.7,and visible edge increase rate(E)increased by 4.8%.Experimental results show that the proposed image dehazing algorithm is effective.