Reading the land from the sky
Machine learning’s ability to interpret large quantities of data, recognise patterns, and interpret images far more accurately and quickly than humans, are at the centre of the solution's potential. It rapidly produces incredibly detailed, region-scaled land use maps from only a satellite image as initial input. It also gives us the ability to create new types of land classification that may not yet exist or have been produced before for an area, such as maps of greenhouse locations, industrial warehouses, mature trees, specific types of urban and rural open spaces, etc., anywhere in the world, with reasonable accuracy.
We have known for some time that nature-based solutions, when compared with their concrete infrastructure equivalents, can provide a variety of additional benefits, but didn’t have the tools to show what is possible and measure the full value of this alternative. In Shanghai we were able to quickly identify opportunities for the city to maximise its parks, lakes and other green features to reduce flood risk.
The solution has many potential uses – from helping a water utility operator in the UK to make the economic case for investing in green and blue infrastructure, to allowing our landscape architects, economists and ecologists to develop a viable scheme for an “orbital forest” surrounding the city of Tirana, in Albania.