Translator Disclaimer
21 October 2014 An overview of neural network applications for soil moisture retrieval from radar satellite sensors
Author Affiliations +
Frequent and spatially distributed measurements of soil moisture (SMC), at different spatial scales, are advisable for all applications related to the environmental disciplines, such as climatology, meteorology, hydrology and agriculture. Satellite sensors operating in the low part of microwave spectrum are very suitable for this purpose, and their signals can be directly related to the moisture content of the observed surfaces, provided that all the contributions from soil and vegetation to the measured signal are properly accounted for. Among the algorithms used for the retrieval of SMC from both active (i.e. Synthetic Aperture Radar, SAR or real aperture radars) and passive (radiometers) microwave sensors, the artificial neural networks (ANN) represent the best compromise between accuracy and computation speed. ANN based algorithms have been developed at IFAC, and adapted to several radar and radiometric satellite sensors, in order to generate SMC products at different spatial resolutions, varying from hundreds of meters to tens of kilometers. These algorithms, which use the ANN techniques for inverting theoretical and semi-empirical models, such as Advanced Integral Equation (AIEM), Oh models, and Radiative transfer Theory (RTT), have been adapted to the C-band acquisitions from SAR (Envisat/ASAR) and real aperture radar (ASCAT) and to the X-band SAR acquisitions of Cosmo-SkyMed and TerraSAR-X. Moreover, a specific ANN algorithm has also been implemented for the L-band active and passive acquisitions of the incoming SMAP mission. The latter satellite will carry onboard simultaneously one radar and one radiometer operating at the same frequency, but with different spatial resolutions (3 and 40 km, respectively). Large datasets of co-located satellite acquisitions and direct SMC measurements on several test sites located worldwide have been used along with simulations derived from forward electromagnetic models for setting up, training and validating these algorithms. An overall quality assessment of the obtained results in terms of accuracy and computational cost was carried out, and the main advantages and limitations for an operational use of these algorithms have been evaluated.
© (2014) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
E. Santi, S. Paloscia, and S. Pettinato "An overview of neural network applications for soil moisture retrieval from radar satellite sensors", Proc. SPIE 9243, SAR Image Analysis, Modeling, and Techniques XIV, 92430F (21 October 2014);


Back to Top