I. Introduction
Absorption or scattering of light travelling through water droplets floating in the air produces haze. The aim of de-hazing methods is to repair hazed image into a clear haze free image, because various progressive image processing need a haze free image [1] – [6] . Methods for hazy image processing can be categorized as: One is enhancement method and second is based on physical model. Enhancement based method solve the purpose of clarity. When there are more differences in scene depth and the image scene needs to be de-hazed, this method does not perform well. Enhancement techniques are Bi-histogram modification [7] , Histogram modification [8] Histogram Equalization [9] , Weighted Histograms [10] , etc. These techniques suffer from saturation of pixels like over and under enhancement. Physical model based algorithms [10] , [11] are used for precisely approximation of global atmospheric light α and transmission map t(x). And by using model of atmospheric scattering, clear images can be recovered. On complicated scenes, this model has a better effect, and image information is saved more thoroughly. The atmospheric scattering model defined as [12] – [13] : \begin{equation*}f(x) = I(x)\,t(x) + \alpha \,(1 - t(x))\tag{1}\end{equation*}