Single Image Haze Removal Using White Balancing and Saliency Map

AbstractAtmospheric phenomenon such as haze, fog, mist etc is originated by small particles in the air such as dust, smoke or moisture, which reduces the visibility of the outdoor scenes. The background theory behind it is the absorption and scattering of light, that passes through these atmospheric particles. In this paper, we instigate a single image dehazing approach which enhances the visibility of the hazy images. The proposed method focus on contrast based single image dehazing. This approach makes use of white balancing, which eliminates the color cast that is caused by the atmospheric color. In this method, we compute the saliency map of the white balanced image, in order to get the well defined boundaries and uniformly highlighted salient region. Morphological opening is performed on the saliency map, to preserve the size and structure of the objects in the image. At the final step we apply contrast enhancing approach CLAHE, to intensify the local contrast of the images. The performance analysis is done on various natural hazy images. This method has been evaluated with other dehazing algorithms reported in the literature. The experimental results show that the method we developed has the capability to remove the haze efficiently and it can retain the fine details of the regions closer to the horizon