Scientists from the Space Research Center of the Polish Academy of Sciences (CBK PAN) have developed a method for automatically and quickly creating detailed maps globally. The prototype of the map, covering only Europe, was commissioned by the European Space Agency.

The map is presenting the relief of Europe in 2017 with 13 most important classes, including agricultural areas, forests (separately deciduous and coniferous), built areas, marshes, peat bogs. As many as 15,000 images from Sentinel-2 satellites were used to develop the map. The smallest distinguishable details are 10 meters of size in reality, which means that the data is up to 2.5 thousand times more detailed than the most commonly used map of this type (CORINE). Thanks to the use of fully automatic algorithms, the production of the new map took only 6 weeks.

The terrain classification method was developed by scientists from the Earth Observation Department of the Space Research Center of the Polish Academy of Sciences. In their work, they used an approach called ‘machine learning’. Researchers tested their method in China, Columbia, Namibia, Germany and Italy, and used it to generate a new terrain map in Europe.

Project website:

http://s2glc.cbk.waw.pl/