g., a Synthetic Aperture Radar: SAR) aboard satellites or aircrafts (e.g., [8]). They are based on experimental campaigns in which the backscattering http://www.selleckchem.com/products/Imatinib(STI571).html radar measurements are coupled with data representing the surface characteristics. The database formed Inhibitors,Modulators,Libraries by the data acquired during the campaigns can be analyzed by means Inhibitors,Modulators,Libraries of a regression approach to derive a relationship yielding the sensor measurement as a function of the sensor characteristics (frequency, observation angle, polarization) and of some quantities representing the soil conditions, usually expressed in terms of dielectric (e.g., soil moisture) and roughness (standard deviation of heights and correlation length) parameters. The regression analysis permits deriving a simple formula, so that the advantage of semiempirical models in terms of simplicity is evident with respect to physically-based approaches.

A critical point is the representativeness of the experimental database, i.e., its ability to encompass a wide set of soil conditions, thus ensuring a large range of applicability of the derived relationship [9].From the previous Inhibitors,Modulators,Libraries discussion, the need to join the simplicity and the efficiency of the semiempirical backscattering models to the Inhibitors,Modulators,Libraries precision of physical ones clearly emerges. To succeed in combining these two key features, a neural network approach can be attempted. Since a multilayer feed-forward neural network (NN), having at least one hidden layer, can approximate any nonlinear function relating inputs to outputs [10], it can be profitably adopted to emulate a forward electromagnetic model giving advantages in terms of computational speed and maintaining a fairly good degree of accuracy.

The adoption of a NN technique to improve the efficiency of forward models Brefeldin_A was applied in [11,12] in order to approximate sea surface scattering models.In this work, a neural network approach to the problem of reproducing the behavior of the IEMM is proposed. We have considered only the backscattering case, because the radar sensors presently operative are monostatic systems, although bistatic experiments have been recently envisaged (e.g., [13,14]). We have made reference to two sensors with different characteristics. The first one is a SAR operating at C-band (5.3 GHz) with an incidence angle ��I = 23��, such as ERS-2 and also ENVISAT/ASAR in some of its acquisition modes.

The second radar configuration, is an L-band (1.25 GHz) instrument with ��i = 34��, similar to ALOS/PALSAR in fine beam modes. We have firstly built two training sets and two etc test databases (one for each frequency band) consisting of matched pairs of vectors of input soil parameters (i.e., soil moisture mv, standard deviation of heights s and correlation length l) and IEMM outputs (i.e., backscattering coefficients denoted as ��0). The incidence angles previously mentioned (hereafter denoted also as nominal incidence angles) have been considered in this case.