Image defogging method combining light field depth estimation and dark channel

In order to improve the defogging effect of fogged images, the traditional dark channel prior theory is easy to cause halo effect in abrupt edge region and color distortion in sky area. In this paper, an image defogging method combining light depth estimation with dark channel theory is proposed. Firstly, block and point transmittance estimation of dark channel are used for multiscale fusion. Secondly, the final transmittance is obtained by combining the transmittance calculated from the depth information of light scene and the transmittance of dark channel. Finally, the sky region is segmented by using the depth information of light scene, and the atmospheric light is obtained in the sky area. Experimental results show that, compared with the traditional de fogging algorithm, the proposed algorithm is optimal in both subjective and objective evaluation.