Cambridge University researchers are deploying a novel system, Tessera, that marries satellite imagery with artificial intelligence (AI) to map hedgehog habitats and understand their precipitous decline across the United Kingdom. The initiative, revealed in reports from the past day and in late September 2025, aims to pinpoint not only where the spiky mammals reside but also the obstacles that hinder their access to food and potential mates.
The core of this project involves an AI model trained to identify likely hedgehog habitats by detecting features like bramble thickets, which are crucial for their survival. This approach offers a more efficient, large-scale alternative to traditional, labor-intensive survey methods. The system analyzes images from the European Space Agency's Sentinel satellites, processing them through machine-learning algorithms to cover vast areas simultaneously.
Habitat Mapping and Environmental Change
Tessera's capabilities extend beyond simply locating potential hedgehog havens. Researchers state that the system's outputs can track the influence of new housing developments and other environmental shifts on the landscape, assessing their long-term impact on hedgehog populations. This allows for a more comprehensive understanding of how human activity shapes the environment critical for these animals.
Read More: Forest Gaps Help New Oak Trees Grow Better, Study Says
The AI model itself is described as utilizing relatively straightforward machine-learning techniques, specifically a combination of logistic regression and k-nearest neighbors classification, rather than large language models. The accuracy of the AI's predictions has been verified through ground-truthing exercises in Cambridge, where researchers compared the model's identified bramble patches against actual terrain.
Broader Conservation Applications and Previous Efforts
While the immediate focus is on hedgehogs, the Tessera system is an open-source project with broader potential applications. Researchers envision it shedding light on various ecological concerns. Insights gained can also be integrated with other data sources, such as information from tiny GPS trackers attached to individual hedgehogs for real-time movement monitoring.
This effort aligns with wider national initiatives, such as the National Hedgehog Monitoring Programme, which has been trialling a combination of trail cameras, AI, and volunteer efforts since its launch earlier in 2024. This programme seeks to establish robust population estimates for hedgehogs, which are known to be in sharp decline, with estimates suggesting a drop between 30 and 50 percent in recent years. The urgency for effective monitoring is underscored by this significant population loss.
Read More: MatterChat AI helps scientists find new materials on 19 May 2026
Technological Underpinnings and Data Integration
The project leverages data from the European Space Agency's Sentinel satellites. This raw imagery is processed using the Tessera system, which integrates what are termed 'earth representation embeddings'. These are then combined with ground-truth observations, including data from citizen science platforms like iNaturalist, to refine the AI's accuracy. The researchers involved in developing this bramble-detecting model include Gabriel Mahler, alongside colleagues Sadiq Jaffer, Anil Madhavapeddy, and Shane Weisz.
