A regularized optimization approach to fast image dehazing

This paper presents a novel and fast linear regularized optimization algorithm for single image dehazing, which is based on two statistical observations. One observation is that images under hazy conditions usually exhibit low contrast and the other is that the spatial distribution of distances from scene objects to the camera is piece wise smooth. In addition to the linear optimization, digital matting and wavelet decomposition techniques are also applied to refine the dehazing results and speed up the computation. Simulations and evaluations of the proposed algorithm in comparison with state of the art algorithms are carried out. The obtained results can demonstrate advantages of the proposed algorithm in color fidelity and recovery of image details.