Akademik Veri Yönetim Sistemi
3rd International Conference on IoT Based Control Networks and Intelligent Systems, ICICNIS 2022, Kottayam, Hindistan, 1 - 02 Temmuz 2022, cilt.528, ss.743-756
Fog and haze naturally or artificially appearing in the environment, limit human visibility. As a way of improving the visibility, digital images are captured and many different image enhancement methods are applied to remove the fog and haze effects. One of the fundamental methods is the Dark Channel Prior (DCP) method. DCP can remove fog and haze on a single image by modelling the psychical diminishing structure of fog. In this study, the spectral signature of the DCP method was investigated by using the transmission maps produced by the results of the DCP method on the Spectral Hazy Image Database (SHIA) dataset, which consists of hyperspectral images taken in the visible and near infrared band range. In this study, it was observed that the transmission response of different regions in the image to the increase in fog density was different. By using this distinctiveness on two foggy images taken from the scene in two different high fog level, this study achieves to reveal the silhouette of the scene which is totally not visible to human eye.