Image dehazing algorithm based on atmosphere scatters approximation model

Abstract

Due to the scattered light of suspended particles in the atmosphere, the images taken in the foggy day become gray and are lack of visibility. In order to unveil the clear images structures and colors, the author propose an algorithm based on atmosphere scatters approximation model, which adopts the extended Jones Matrix and Stokes Law to calculate approximate the transmission of light in the atmosphere, so as to eliminate some of the scattered light. Both the light intensity in the atmosphere and haze concentration are obtained by means of Dark Channel Prior, afterward the extinction function for light transmission is used for calculation to restore the foggy images. The experimental results show that the algorithm can not only effectively improve scenery visual effect under different condition of haze, and provide clear pictures for machine vision applications in the foggy day. © 2012 Springer-Verlag.