The main objective of this paper is to assess the capability of a soil moisture (SMC) algorithm adapted to the GMES Sentinel-1 characteristics, developed within the framework of an ESA project (SMAD-1). The SMC product shall be generated from Sentinel-1 data in near-real-time and delivered to the GMES services within 3 hours from observations. Two different complementary approaches were proposed: the first approach was based on Artificial Neural Networks (ANN), which represented the best compromise between retrieval accuracy and processing time, thus being compliant with the timeliness requirements. The second approach was based on a Bayesian Multi-temporal method, allowing an increase of the retrieval accuracy, especially in case of few ancillary data available, at the cost of computational efficiency, taking advantage of the frequent revisit time achieved by Sentinel-1. The algorithm was validated in several test areas in Italy, US and Australia, and finally in Spain by performing a 'blind' validation. © 2012 IEEE.
An algorithm for soil moisture mapping in view of coming Sentinel-1 satellite / S., Paloscia; S., Pettinato; E., Santi; Pierdicca, Nazzareno; Pulvirenti, Luca; C., Notarnicola; G., Pace; A., Reppucci. - STAMPA. - (2012), pp. 7023-7026. (Intervento presentato al convegno 2012 32nd IEEE International Geoscience and Remote Sensing Symposium, IGARSS 2012 tenutosi a Munich nel 22 July 2012 through 27 July 2012) [10.1109/igarss.2012.6351954].
An algorithm for soil moisture mapping in view of coming Sentinel-1 satellite
PIERDICCA, Nazzareno;PULVIRENTI, Luca;
2012
Abstract
The main objective of this paper is to assess the capability of a soil moisture (SMC) algorithm adapted to the GMES Sentinel-1 characteristics, developed within the framework of an ESA project (SMAD-1). The SMC product shall be generated from Sentinel-1 data in near-real-time and delivered to the GMES services within 3 hours from observations. Two different complementary approaches were proposed: the first approach was based on Artificial Neural Networks (ANN), which represented the best compromise between retrieval accuracy and processing time, thus being compliant with the timeliness requirements. The second approach was based on a Bayesian Multi-temporal method, allowing an increase of the retrieval accuracy, especially in case of few ancillary data available, at the cost of computational efficiency, taking advantage of the frequent revisit time achieved by Sentinel-1. The algorithm was validated in several test areas in Italy, US and Australia, and finally in Spain by performing a 'blind' validation. © 2012 IEEE.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
