Applicazione di reti neurali a dati SSM/I multi-temporali per la stima di parametri del manto nevoso

In this paper; different configurations of art$cial neural networh are proposed and analysed for the extraction of snow parameters from radiometric remotely-sensed data. Brightness temperatures collected at 19 and 3 7 GHz by the SSM/I are used to retrieve snow depth, fiactional volume, temperature and mean particle size. Artificial neural networks are trained either with experimental data and with brighntess temperatures simulated by means of a mode1 based on the Dense Medium Theory (DMRT). The area selected for our study is located in North Europe by using data collected between the years 1996 and 1999.