A Mixed Transmission Estimation Iterative Method for Single Image Dehazing

Considering the frequent occurrence of hazy weather, single image dehazing has become an important research task. Physical model-based and deep learning-based methods are two competitive in dehazing, but achieving fidelity effectively removing haze at same time real scenes is still a challenging problem. In this work, we propose mixed iterative model to restore high-quality clear images by integrating physical approach approach, which can combine advantages method maintaining natural attributes completely haze. For image, first, according density, it divided into separate regions calculate local atmospheric light. Then, utilize dark channel prior DehazeNet jointly estimate transmission, accurate recovered that more line with scene. Finally, numerical strategy employed further optimize light transmission. The experiments on both synthetic datasets indicate proposed clean superior performance than other state-of-the-art algorithms peak signal-to-noise ratio, structural similarity, information entropy, color index, histogram. And four indicators quantitative results test set, achieve highest score three them second result another. comparison images, our average objective highest.