A Fast Advanced Algorithm for Haze Removal by using Color Attenuation Prior

Abstract

The problem of haze removal has a long and well-travelled history. Dehazing from a single input hazy image is very challenging. This is because we have a little knowledge about the image. Since concentration of the haze in a hazy image is different from place to place it will become very hard to detect haze in image. The proposed work is a fast single image haze removal technique and it is work by using color attenuation prior. This method remove haze progressively. A linear model is created and the depth map is produced by using that model. From depth map estimate the transmission diagram and restore the scene radiance. Thus haze can be removed efficiently. Experimental results show that color attenuation prior model will work more efficiently than state-of-art haze removal algorithms.