Semantic Integration of Raster Data for Earth Observation on Territorial Units

Semantic technologies have proven their relevance in facilitating the interpretation of Earth Observation (EO) data through formats such as RDF and reusable models, especially for representation space time. While rasters are usual format results image processing algorithms, a recurrent problem is transferring pixel values these into features that make sense areas interest on Earth, thus content. This paper addresses this issue semantic integration process based spatial temporal properties. We propose (i) modular generic model homogeneous qualifying geographical area thanks to territorial units (land parcels, administrative units, forest areas, etc.) we define divisions larger territory according criteria relation with human activities; (ii) extraction, transformation load (ETL) builds extracted from maps aggregated corresponding unit areas. evaluate our approach terms adaptability proposed pipeline accommodate different use cases (vineyard urban expansion monitoring), added value generated datasets assist decision making, (iii) scalability approach.