Accuracy Assessment of a 122 Classes Land Cover Map Based on Sentinel-2, Landsat 8 and Deimos-1 Images and Ancillary Data

Castile and Leon crops and natural land map is a highly detailed land cover layer, obtained through satellite imagery, which distinguishes between 122 land cover classes and includes 50 specific crop types. The project began in 2013 by using several satellites, with the production cost greatly reduced since 2016 when Sentinel-2 imagery became freely available, and is updated annually. The classification is performed using a machine learning algorithm trained with data retrieved from Integrated Administration and Control System and some other land use databases available in Spain. This map is also proposed as an advanced crop map, within SENSAGRI project (Sentinels Synergy for Agriculture) drafting in response of the EO Work programme “EO-3-2016: Evolution of Copernicus Services”, as one of the four advanced proof-of-concept services. The algorithm will be validated in others European agricultural test areas which, along with Castile and León, are representative enough to show an overview of the European crop diversity.