Multispectral Satellite Image Dehazing and Pollution Estimation

Abstract

Since the introduction of computer vision, the difficulty of image dehazing exist. Due to the presence of numerous air particles, which causes hazed photos, images taken during bad weather circumstances sometimes appear to be of inferior quality. To eliminate haze from the photos, this study uses a static technique called dark channel before. Ground based air pollution sensors give regular readings but have limited coverage. Satellites offer broad global coverage but poor spatial resolution and insufficient data on the vertical distribution of pollutants. It is not easy to estimate the concentrations of pollutants at the surface, where they are produced. In order to address this problem, a framework has been developed. This research work focuses on providing an approach incorporating the dehazing of a multispectral satellite picture and providing a pollution range of the region by estimating the percentage of nitrogen oxides.