CLCplus Backbone vector 2018 is a land cover map that contains vector polygon geometries (minimum mapping unit: 0.5. ha; minimum mapping width: 20 m) and is based on image segmentation techniques of Sentinel satellite time series (Softbone) and a combination of existing reference datasets for geometries containing transportation and hydrological networks (Hardbone). Each polygon represents aggregated landscape objects and contains their dominant land cover among the 18 basic land cover classes. In addition, polygons are enriched with land cover fractions from the CLCplus Backbone raster as well as aggregated attributes based on other CLMS and Copernicus products (e.g. topography).

Input data:

1) Softbone: landscape objects obtained from image segmentation techniques of Sentinel-2 L2A data acquired from 2018-04-01 to 2018-10-31.

2) Hardbone: persistent linear landscape objects such as transportation network (i.e. roads, railways) and hydrological waterways (i.e. rivers, lakes, canals) from several external sources (e.g. Open Street Map, CLMS EU-Hydro river and canal network, Water Information System for Europe (WISE) surface waterbody lines, national/regional road and hydrology data).

Processing and Attributes:

1) Landscape objects are generated by combining two levels of objects - level 1 and level 2 objects - so called “Hardbone” and “Softbone”, respectively. Hardbone represents a geometric skeleton of persistent linear landscape objects such as transportation network (i.e. roads, railways) and hydrological waterways (i.e. rivers, lakes, canals), which comprises of various line vector data. Softbone further delineates landscape objects in a controlled automatic manner based on the image segmentation technique, which generates landscape object polygons by their spectral response and variation throughout a year within the frame of Hardbone.

2) Each polygon is assigned with its dominant land cover class (based on majority rule) and the areal fractions (in percent of the total area of the respective polygon) of the three most abundant underlying CLCplus Backbone raster 2018 classes. In addition, CLCplus Backbone vector 2018 is enriched with 58 attributes derived from other CLMS and Copernicus products: CLMS Urban Atlas, CLMS Riparian Zones, CLMS Coastal Zones, CLMS Tree Cover Density, CLMS Imperviousness Density, CLMS Water and Wetness, CLMS High-Resolution Vegetation Phenology and Productivity (HR-VPP), CEMS European Forest Fire Information System (EFFIS) and Copernicus DEM 10m.

Postprocessing:

Steps applied comprise 1) smoothing of Softbone geometries, 2) reintroduction of Hardbone lines, 3) snapping vertices, 4) cleaning the remaining MMU errors and topological consistency (gaps and overlaps).

Changes to previous release:

Not available