An algorithm based on neural networks for generating multi-temporal soil moisture maps from ENVISAT/ASAR images

In this paper the actual capabilities of ENVISAT/ASAR images in providing soil moisture maps have been tested. Several SAR images were collected on two test areas: a flat agricultural region in the Alessandria area, in Italy, and the natural area of Kemijoki river system, in Finland. An inversion algorithm based on Artificial Neural Networks (ANN) for the retrieval 4-5 levels of soil moisture from backscattering data was tested and successfully compared to ground measurements.