Improving Single Image Haze Removal Based On Cellular Automata Model

Hazy images are the images acquired under bad weather with low contrast and faint color properties. Several algorithms used dichromatic model to remove the haze in the image. This paper proposed a novel technique using cellular automata for the single image haze removal. It aims to improve the dark channel and the transmission map. We proposed the cellular automata rule to refine the intensity of image pixel in a dark channel. Then, the light source is estimated from the dark channel. Finally, we proposed the cellular automata rule to construct the transmission map and restored the haze-free image. The experimental results show that the proposed method improved intensity, color saturation quality, and avoid halo artifact without any post-processing when compared with the state-of-the-art methods.