INVESTIGATION ON THE CLASS DEPENDENT PROPERTIES OF ATMOSPHERIC SCATTERING EFFECTS ON SATELLITE SIGNAL IN THE VISIBLE AND NEAR INFRARED REGION OF SOLAR SPECTRUM

摘要

Atmospheric scattering is one of the key factors affecting the precision of remote sensing retrieval of geo-information from satellite based radiative measurements. In the present paper, an investigation has been carried to characterize and quantify the effects of atmospheric scattering on the satellite signal corresponding to visible and near-infrared region of the solar spectrum. A coupled surface-atmosphere radiative transfer model has been employed for the purpose. Radiative responses at the top of the atmosphere have been generated using coupled model under different atmospheric and geometric condition for different surface reflectance values covering soil, vegetation and water. The study depicts that radiative contamination due to atmospheric effects very much dependent on the surface classes particularly on the pixel-based response characteristics. This condition imposes a serious restriction on the widely used method of dark pixel correction which can obtain erroneous results depending on the amplitude of surface reflectance of radiative responses. Radiative transfer simulation shows that the amplitude can be either increased or decreased for the identical atmospheric configuration. Such a situation suggests utilization of a specific radiative correction tools for precise and accurate interpretation of the signal. In this connection, effects of atmospheric optical depth on the satellite image characteristics have been studied using a Landsat TM image-derived spectral reflectance coupled with an atmospheric model so as to simulate images at the top of the atmosphere (TOA) for different atmospheric aerosol contents.