一种改进的有雾图像去雾算法

针对基于大气散射模型的图像去雾算法存在的图像去雾后颜色偏暗、对比度过度增强的问题,提出一种基于灰色关联度引导滤波的图像去雾算法。首先,假设有雾图像的像素可以分为正常像素和被雾霾颗粒破坏的像素,应用灰色关联理论对雾霾图像的像素值进行判断;然后,对雾霾颗粒破坏的像素进行引导滤波,通过取对数的方法缩小原始图像和滤波以后图像像素值之间的差异,在对数域中计算雾气面纱值;最后,依据大气散射退化模型反演复原清晰的图像。实验结果表明,该算法不仅可以有效改善雾霾图像的清晰度,而且能够解决去雾后存在的亮度偏暗,色彩失真等问题。

An improved foggy image dehazing algorithm

Aiming at the problems that there is dark color and excessive contrast enhancement after image defogging existed in image defogging algorithm based on atmospheric scattering model, a fast image defogging algorithm based on gray correlation degree guided filtering is proposed. Firstly, assuming that the pixels of foggy images could be divided into normal pixels and pixels damaged by haze particles, the gray correlation theory is applied to judge the pixel values of haze images. Then, the pixels damaged by smog particles are guided to filter, the difference between the pixel values of the original image and the filtered image is reduced by taking logarithm, and the fog veil value is calculated in the logarithm domain; Finally, a clear image is retrieved based on the atmospheric scattering degradation model. While Experimental results show that the proposed algorithm could not only effectively improve the clarity of haze images, but also solve the problems of dim brightness and color distortion after defogging.