单幅图像快速去雾算法

目前去雾算法主要有通过暗原色和对图像颜色通道处理等方法，但是这些方法去雾效率不高，从而导致实用性不强，针对此弊端提出了一种基于单幅图像的快速去雾算法。大气光估计运用改进的暗通道方法，先对颜色通道进行最小滤波，然后取最小滤波的最大值作为大气光的估计值；透射率估计运用物理模型均值滤波，先根据数学模型转换，然后进行一次均值滤波，再用偏移值来修正带透射率的估计值。算法简单快速有效，具有实用性。对实验结果进行定性定量分析，证明与其他算法相比，所提算法具有更好的去雾效果和更快的处理速度。

Fast Single Image Defogging Algorithm

At present, the defogging algorithm mainly includes the dark channel prior method and the image color channel processing method, but the defogging efficiency of these methods is not high, resulting in poor practicability. In view of this drawback, a fast defogging algorithm based on single image is proposed. Atmospheric light estimation uses the improved dark-channel approach, at first make color channel minimum filtering, and then take the maximum value of minimum filtering as atmospheric light estimation. Transmittance estimation uses physical model mean filter, first convert according to the mathematical model and a mean filter, then use the offset to correct the estimated value with transmittance. The algorithm is simple, quick and effective, and has practicality. Qualitative and quantitative analyses of the experimental results show that compared with other algorithms, it has better defogging effect and faster processing speed.