United KingdomPublic sector2016 – ongoingNatural England
Living England machine-learning habitat classification
Natural England produced national habitat predictions from satellite imagery and field observations
Problem
National habitat surveys were expensive and difficult to repeat. Natural England needed maps that could be updated across the whole country.
Approach
Living England used random forests from the project's start in 2016. Its in-house team combined habitat labels from field observations with satellite imagery and environmental variables. The production process generated habitat predictions with reliability scores for policy and land-management use.
Results
No published figures
Takeaway
The habitat map remained a predictive baseline. Natural England required users to combine it with other evidence and local land advice.
Categories
- Industry
- Public sector
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and MLComputer vision
- Adoption
- Built in-house
- Evidence
- Company's own claim
Sources
Source summaries
Natural England described its production random-forest habitat mapping process and its development since 2016. It explained how model reliability was assessed and how the predictions should inform land-use decisions.
Full details are in the original sources.