Image Defogging Algorithm Based on Fisher Criterion Function and Dark Channel Prior

Abstract

For images with sky region, we propose a method of image defogging based on pattern recognition and dark channel priori. First, the transmittance is obtained by reducing the minimum filter window, and then the edge of transmittance is smoothed by V transformation. Since the transmittance histogram has the feature of two peaks, we used Fisher linear classifier to identify sky and non-sky regions, and then recalculated the transmittance of sky regions. Finally, the atmospheric scattering model is used to get the new luminance component, and the gamma correction is carried out. After the saturation component is adjusted appropriately, the defogged image is converted back to RGB space. The experiment shows that the proposed algorithm can get better haze removal effect compared with the traditional defogging algorithm based on dark channel prior and other state-of-the-art methods developed in recent years, and to a certain extent has solved the problem of color distortion and the whole brightness being darker in the sky region when using the DCP based algorithm.