Single Image Dehazing using Adaptive Gamma Correction Method

In this paper, an improved single image dehazing method is proposed to remove the haze effect. The proposed algorithm used a very simple and effective approach for haze removal. Many haze removal algorithms have been developed, but halo artifacts and color distortion problem persists. The dark channel prior (DCP) method is the best method for single image dehazing. But, it is applicable only for the small sky region. The dark channel prior (DCP) method still faces the color distortion problem due to the large sky region present in input hazy image. So, we proposed an effective approach that removes the haze from a single input image using the adaptive gamma correction method. In this proposed algorithm, first, we segment the large sky region using the k-means clustering approach and then refined the whole segmented sky region. Here, a guided filter method is used for the refinement of the segmented sky region. It provides edge related information accurately. In most of the dehazed methods, the output haze-free image suffers the brightness level or color distortion problem. So, in the proposed approach gamma correction method is applied to the filtered image. It reduced the color distortion from the hazy image properly.