Retrieving visibility values using satellite remote sensing data

The recent development of satellite meteorology has allowed us to estimate spatially and frequently number of basic meteorological parameters. This paper presents the proposed methodology for retrieving visibility values based on the application of the darkest pixel atmospheric correction algorithm on satellite image data. The method is based on the use of the radiative transfer calculations followed by some key assumptions. Landsat-5 TM band 1 images (0.45–0.52 μm) have been used to determine the visibility value for each image date. A direct comparison between the measured visibility data from the airport meteorological stations with the determined visibility data was performed showing high correlation values. Indeed, by relating the determined visibility data with those measured on the Heathrow Airport station in the West London (UK), a correlation coefficient of r2 = 0.97 has been found with the observed significance for the regression model to be less than 0.05, for four multi-temporal images acquired on 1985 and 1986. The algorithm has been tested also to Landsat TM images of the Paphos Airport area in Cyprus with satisfactory agreement between the visibilities measured at the meteorological station and those found from the images. The algorithm presented may be useful for assessing the atmospheric conditions of satellite images and also can assist the improvement and effectiveness of the available atmospheric correction algorithms.