Modelo de la geometría sol - sensor en la reflectancia de la vegetación: validación a nivel satelital

The observations of the terrestrial surface by means of satellite remote sensing, have the potential of being used to characterize processes occurring on the soil-vegetation-atmosphere system. Data acquired from different view angles are an additional source of information with respect to those offered for spectral, spatial and temporal variations detected with this kind of instruments. One of the applications where the use of this fourth source of information has become relevant is land cover classification. To explore its usefulness, a number of approaches have been proposed based on linear kernel-driven models, unfortunately no significant contribution on the associated parameters of the models to classify land cover has been reported. On this work two issues are investigated: (1) the fitness of the Uni-Parametric BRDF model to data acquired by the POLDER-1 sensor and (2) the contribution of multiple view angle observations to the land cover classification problem using the aforementioned model. Although correlation levels between the measurements of the POLDER-1 sensor and those by the model under consideration are greater than 97%, no evidence was found related to the improvement of land cover classification by using multiple view angle data. These results are in agreement with the need to define new classification systems based on variables that may be inferred through the observations acquired by a satellite remote sensors reported by other researchers.