基于强度反转和变分Retinex模型的图像去雾方法

A Hazy Image Enhancement Method Based on Intensity Inversion and the Variational Retinex Model

提出一种基于强度反转变分Retinex模型的雾天图像增强方法。首先,通过反转操作将雾天图像转化为虚拟暗图像,并基于变分Retinex模型快速估计出虚拟暗图像所对应的入射分量图。然后,对雾天图像中的天空区域进行识别,并在此基础上对入射分量图进行修正,进而获取对应的反射分量图。最后,通过对反射分量图进行反转操作来获得消除雾气后的清晰图像,并引入一种图像亮度调整算法进一步提高清晰图像的视觉效果。实验结果证实了所提方法的有效性。

A hazy image enhancement method based on intensity inversion and variational Retinex model was proposed.Firstly,a virtual dark image was obtained via hazy image inverting,and the corresponding illumination component of the virtual dark image was estimated based on the variational Retinex model with high-efficiency.Then,the sky region recognition within the hazy image was realized,an illumination component map refinement method was presented accordingly,and therefore the corresponding reflection component map can be obtained.Finally,the clear image after haze-removal was obtained by inverting the reflection component map,and an image brightness adjustment method was utilized to further improve the visibility of the clear image.Experimental results verify the effectiveness of the proposed method.