The limited visibility caused by fog and haze is a most important problem for several function therefore haze removal by these images for visibility improvement is necessary. In this paper, haze removal is done by make use of the dark channel prior procedure and estimation of the Atmospheric light technique. Dark channel region is calculated by using dark channel algorithm then the atmospheric variation is calculated on that dark channel region. The estimate of pixel value in the dark region and the atmospheric variation estimate is the estimate of transmission obtained as the transient image. Further the optimization of data is done by smoothing to obtain the radiant image or the dehazed image