On the Performance analysis of image Dehazing using fuzzy theory and Artificial Neural Networks

Abstract

Photography in hazy environment, light attenuation and scattering caused by the water particles present in the medium, result in loss of severe image quality and loss of valuable information. In order to minimize the effect of haze and improve visual quality, this literature present a novel technique combining fuzzy theory, artificial neural networks and image fusion. Transmission map is estimated using fuzzy inference system. Then morphological operation and artificial neural network are applied to remove the halation present. Backpropagation, feedforward, cascaded-feedforward and fitnet artificial neural networks are applied on halation free transmission map for further refinement. Finally, image fusion technique is used to recover an enhanced version of all four images.