Image Dehazing Based on Robust Sparse Representation

Abstract

One source of difficulties when processing outdoor images is the presence of haze, smoke, or fog, which reduces the contrast and fades the colors of the observed object. To overcome this problem, we propose a novel single-image dehazing method based on robust sparse representation in this paper. First, we propose to employ the robust sparse representation on the color and luminance channels of the hazy image to obtain the haze density information. Then, we introduce the haze density information to estimate the transmission. Using the method based on the atmospheric scattering model, we can directly recover a natural haze-free image. Finally, in order to improve the visual quality of the dehazed image, a post-processing procedure is designed to enhance the visibility. We conduct the experiments on diverse types of hazy images and the experimental results show that the proposed method based on robust sparse representation outperforms several traditional and state-of-the-art methods.