Image of the outdoor scene is usually degraded due to poor weather conditions,thus forming a hazy image.So far,most single image dehazing algorithms ignore the effect of noise.Therefore,in consideration of the noise pollution,a new dehazing model is established,and a three-stage dehazing algorithm is proposed.In the first stage,the degraded image is preprocessed,to eliminate the interference of noise with image.In the second stage,a Median Dark Channel Prior (MDCP) algorithm is used for restoring the preprocessed image in spatial domain,enhancing the global contrast of the image.In the third stage,the local contrast of the image is enhanced in transform domain,for further restoring the details of image.Compared with those methods using the spatial domain only or without taking into consideration of the noise effect,the proposed algorithm can significantly improve the image quality and definition.