Land Cover London – Satellite Land Cover data highlighting and measuring urban areas, west London. Based on an adapted version of the CORINE land cover schema (© 4 Earth Intelligence and Open Street Maps, 2020)
Earth observation company 4 Earth Intelligence (4EI) has launched a new land cover mapping service based on advanced machine learning. Created from satellite imagery the 4EI service can automatically provide large area base maps together with regular updates of land cover change. Offering a better understanding of changing landscape and vegetation patterns the 4EI Land Cover data also provides insight into the interaction between human activity and nature including improved understanding of the importance of green infrastructure – essential ingredients for solving urban and climatic challenges.
“Land Cover mapping has many applications and can inform decision making at different levels,” commented Donna Lyndsay, Commercial Director at 4 Earth Intelligence. “With standard classifications ranging from ‘artificial surfaces’, ‘agricultural areas’ and ‘water bodies’ all the way down to specific categories such as ‘vineyards’, ‘airports’ and ‘peat bogs’, it is very easy to get an understanding of current land use and how this is changing.”
Applications of 4EI’s Satellite Derived Land Cover data include the identification of habitats which could be vulnerable due to urban sprawl, understanding and demonstrating compliance with planning policies, creation of corporate mitigation strategies and evidence of Corporate Social Responsibility (CSR) objectives. 4EI has already worked with government and commercial organisations around the world to deliver base line maps and updates.
Green Space London – Satellite Land Cover data highlighting and measuring green spaces, west London (© 4 Earth Intelligence and Open Street Maps, 2020)
“Satellite imagery as a source for wide scale land cover mapping offers many advantages,” continued Lyndsay. “Each individual satellite footprint can cover an area as large as 10,000 kilometres squared and outputs can be repeated at regular intervals, in some cases as frequently as monthly. It is these characteristics that contribute to a robust, repeatable and automated production methodology delivering results that are both consistent and objective.”
In order to produce Satellite Derived Land Cover data 4EI can consider individual satellite images or mosaics of images that are close in date. Before classification different indexes, for example NDVI (normalized difference vegetation index) and SAVI (soil-adjusted vegetation index) are automatically calculated and stacked up along with the spectral bands from the satellite images. This is then processed using machine learnt algorithms before post classification quality control is undertaken.
The 4EI schema used to classify land use is adapted from the CORINE (Coordination of Information on the Environment) programme. Originally initiated by the European Commission and latterly administrated by the European Environment Agency the CORINE land cover project defines 44 classifications of land cover and presents results as a cartographic product.