Sentinel-1 SAR data for mapping agricultural crops not dominated by volume scattering

In this paper, a C-band SAR classification algorithm mapping agricultural crops dominated/non-dominated by volume scattering is described and assessed. The algorithm exploits cross-polarized SAR data and it is a part of the SMOSAR algorithm (“Soil MOisture retrieval from multi-temporal SAR data”) developed in view of the forthcoming Sentinel-1 data. The performances of the classification algorithm have been assessed on RADARSAT-2 data acquired over Flevoland (The Netherlands) and Indian Head (Canada) during the ESA AgriSAR´09 campaign. The results indicate that the selected method is fairly robust versus changes of site location and in average it is expected to achieve an overall accuracy equal or better than 80%.