Problematic: The problematic is related to the retrieval and validation of environmental parameters from remote sensing data. Application is done for soil erosion and hydrology. It is a matter of establishing a statistical correlation model between the studied phenomenon and the satellite measure. The result is obtained either by means of a direct model (for example, estimating the evapotranspiration as a direct function of the temperature gradient coming from thermal imagery) or an inverse model (estimating soil erosion by optimizing the difference between an observed temporal signal and a modelled signal depending on the erosion rate). This work is complementary to the previous one on image processing and physical models and also complementary to theme 2 on inverse problems. This separated section is devoted to explain the context in which it is contemplated: studies are led in collaboration framework with specialists in environment. These application-geared works are the opportunity of obtaining data, running problems analysis, and creating a favorable framework for the effective validation of the entire research.
Objectives:
Studies conducted within the ECOAIR project (Brasil partner) on soil erosion form a basis for the whole set of applications around vegetation and hydrology. These studies will be deepened and extended incorporating our experience coming from the past European project IWRMS (partners in the south part of the african continent) about modelling in hydrology.
The use of active sensors, which offer high potential for these applications, will lead to specific image modeling.
Lastly, these highly applied studies are led in cooperation with specialists. They give rise to validation opportunities which must be systematically exploited, notably to validate the models used in the framework of the image processing and physical models activity.