Satellite remote sensing imagery represents an attractive data source to monitor large regions with frequent updates. In this context, the operational production of accurate land cover maps plays an important role in global-scale environmental assessments and becomes crucial for a wide range of research domains. New earth observation missions such as Sentinel provide images with high spatial and temporal resolution. Accordingly, new image classification methods for the generation of reliable land cover maps are needed. In the framework of the Sentinels Synergy for Agriculture (SENSAGRI) project at \textit Centre d'Études Spatial de la Biosphere (CESBIO) in Toulouse (France), this work aims to describe new schemes for detecting crop areas along the agricultural season. The research has focused on performing statistical fusion at decision-level to combine classification results in order to exploit the synergies between Sentinel-1 and Sentinel-2 image times series.