Natural England clears clouds with machine learning | Computer Weekly

Living England is among the latest batch of artificial intelligence (AI) projects listed as part of the government’s drive to showcase the uses of AI and open data in the public sector.

Run by Natural England, the project uses satellite imagery, field data records and other geospatial data to create a national habitat map of England. Rather than the manual surveys of the past, Living England uses AI to track changes to habitats more efficiently, which the government said will speed up decisions around planning and land use while better protecting nature.

Living England demonstrates how combining multiple datasets with human observations improves the accuracy of machine learning. The datasets include satellite data from Copernicus, the Earth observation component of the European Union’s space programme.

The project, which began in 2016, is funded by the Department for the Environment, Farming and Rural Affairs (Defra) through the Natural Capital and Ecosystem Assessment (NCEA) programme.

The European Space Agency (ESA) and European Union (EU) Copernicus Sentinel-2 programme provides imagery of the Earth’s surface at a spatial resolution varying from 10m to 60m. As an optical satellite, the land surface is often obscured by cloud and its appearance modified by cloud shadow, both of which reduce the quality of the imagery used as input for habitat modelling.

Cloud and cloud shadow need to be removed prior to modelling to produce cloud-free mosaic images of the land surface for use in modelling. To achieve this, data collected by surveyors on the ground is correlated with satellite data from the Sentinel-2 programme using machine learning techniques to identify clouds.

Living England also makes use of the Environment Agency’s National LIDAR Programme, which provides high spatial resolution (1m) topographic data. Other datasets include climatic data from the Met Office’s HadUK gridded climate product, geology and soils datasets from the British Geographical Survey, Cranfield NATMAP soilscapes, and Ordnance Survey datasets.

Our Living England project is harnessing the power of AI to inform and support planning decisions far more efficiently

Commenting on the project, Natural England’s chief scientist, Sallie Bailey, said: “Nature restoration, development and economic growth are not opposing forces – they can and must work together to create a sustainable future for people and wildlife.

“Our Living England project is harnessing the power of AI to inform and support planning decisions far more efficiently. This means we can make the biggest impact for nature recovery while helping to deliver the new homes and infrastructure the country needs.”

The government has published an AI Playbook, with case studies and best practices, which it hopes can offer the public sector guiding principles on how to build AI to help their organisations fix services for citizens and support its ambition to transform public services with AI.

Technology secretary Peter Kyle said: “The publication of our AI Playbook today comes with a call to arms for tech specialists across the public sector – use the guidance we are sharing to put AI to work in your organisations at whiplash speed, so we can repair our broken public services together.”

As well as Living England, the government has provided 13 other examples of how AI and algorithmic tools are used to speed up decision-making and improve public services. These include the use of AI to improve weather predictions and in maintaining high standards at MOT testing centres.