Large-scale land use by remote sensing — batch test zone Scot Rovaltain

CEREMA worked in 2013 as part of an exploratory study for the DREAL Rhônes Alpes, the DRAAF Rhône-Alpes and the Regional Council of Auvergne-Rhône-Alpes (Région Auvergne-Rhône-Alpes) on the semi-automatic implementation of a top-level land use layer on the territory of the Rovaltain SCOT based on satellite images. The objective of the method is to help the departments in charge of planning territories to calculate the quantified indicators characterising land use and its evolution in order to meet the regulatory requirements in this area. The method of constructing the layer consists of combining automatic extraction of information from satellite images (supervised classifications) and post-processed information from existing databases (BD Topo, RPG...). If the construction process differs from the IGN process for setting up the future OCS Grande Scale layer, and if the layer specifications are less ambitious, the objective is to have a layer consistent with the national benchmark: compatible nomenclature, use of the same structural skeleton of the territory, same minimum collection unit etc. The interest of such data lies in the following points: — temporal homogeneity of the source information (the satellite images used — RapidEye images at 5 m of GEOSUD resolutions) are acquired on the national territory a few months apart; — possibility of annual updating, due to the annual update of the national satellite coverage, and the use of automated means; — limited costs (satellite images are available free of charge to institutional users, and process automation reduces the human time required for photo-interpretation methods, for example); — response to a first level of needs (zero status of the “natural frame” and “urban reinforcement”, quantification of the distribution of urban or natural areas, quantification of the effectiveness of different planning measures taken, etc.) the result is in vector form, in 4 soil cover classes (anthropised surfaces, water surfaces, woody vegetation and non-woody vegetation). The smallest polygons make an area of 500 m² The occupation type is stored in the LABEL attribute and takes the following values 0 = no data 11000 = Antthropised surfaces 12000 = water surfaces 21000 = woody vegetation 22000 = non-woody vegetation