Intelligent sensor construction using electromagnetic polarimetric modeling and their exploiting in the region environment interpretation

This paper presents a new approach aiming to build an intelligent sensor exploiting a statistical model, discovered by the application of the radiative transfer theory and its use in the remote sensing domain. its purpose is to interpret inhomogeneous scenes, particularly in mountainous regions covered by snow. First, the statistical model is built by the application of radiative transfer theory on mountainous regions, then, calculating their zero and first order solutions at the boundary conditions. These solutions consist of three types of terms: a surface scattering, a surface-volume scattering and a volume scattering terms. For each term, the scattering energy is computed in order to represent the surface content. So, the energy of the total scattering phenomenon at a pixel serving to detect its exact content. The highest energy value, corresponding with the biggest scattering term, gives an important information about the surface nature. Second, prior knowledge sources are exploited to interpret the same previous scene. Where these sources related to the geometrical and physical models of the scene and the transmitted signal of active radar system image. Finally, to improve the efficiency of the statistical model results, these ones will be compared to the prior knowledge sources results, if both are found the same. We summarize this result by considering this statistical model as a prior knowledge source and moreover, it can be used to design an intelligent sensor to be used in remote sensing applications.