Abstract: A novel single image dehazing method with an adaptive gauss filter is proposed to solve the problem in the traditional method that when the medium scene transmission is estimated under the assumption that neighbouring pixels have similar depth,it often leads to a undesirable shift on the depth edge in image dehazing and serious influence to the dehazing effect.Firstly,to simplify the parameter estimates,an improved atmospheric scattering model is given,the estimation of scene transmission can be equivalent to the estimation of veiling in the derivation of the model,that is refined using an adaptive gauss filter based on the spatial proximity and the depth edge.Then,the airlight in a given hazy image is estimated based on the quad-tree subdivision,thus,a high-quality haze-free image can be recovered.Finally,an experiment over several types of outdoor hazy images demonstrates that the proposed approach outperforms state-of-the-art haze removal methods in the efficiency of recovered haze-free radiances and eliminated halo effect,and a high-accuracy image can also be obtained,which results in a great improvement in image visibility.