Satellites and AI track UK hedgehogs to combat decline

At the University of Cambridge, researchers are employing satellite imagery and artificial intelligence to help reverse the decline of hedgehog numbers in Britain.

According to BBC, with the assistance of the AI tool Tessera, which examines high-resolution images of the UK taken from orbit, specialists can accurately pinpoint hedgehog habitats and identify areas where these are vanishing.

These maps reveal landscapes with extraordinary detail, illustrating even individual hedgerows, and AI can precisely forecast hedgehog-friendly areas hidden by cloud cover.

The team behind this initiative aims to illuminate not only the locations of hedgehog populations across the nation but also to understand the obstacles they encounter while searching for food and partners.

The researchers suggest that Tessera's findings could be crucial for assessing the effects of new housing and environmental shifts on territories that might influence hedgehogs over time.

These findings are also being integrated with data from other sources, such as tiny GPS devices attached to some hedgehogs, to observe their real-time movement.

A related programme aimed at safeguarding hedgehogs with "backpack"-style trackers is already in progress in Northern Ireland.

Yet, the move to utilise AI systems and models capable of processing vast datasets and identifying significant trends is part of a broader international conservation effort.

Nonetheless, some express concern about the energy-intensive technology, worried about its potential environmental repercussions.

'Digital Hedgehogs'

In recent years, hedgehog populations have significantly dwindled across Europe.

Within the UK, a 2022 study estimated up to a 75% drop in their numbers in rural locations since 2000.

The common western European hedgehog, the only indigenous species in the UK according to the RSPCA, is now classified as 'Near Threatened' by the International Union for the Conservation of Nature.

Professor Silviu Petrovan, from the People's Trust for Endangered Species, is hopeful that AI technologies will aid in preserving these small mammals.

"Our expectation is to use these advanced models to discern, for example, the particular challenges hedgehogs face in locating food and mates, as well as moving safely in rural areas," he explains.

He mentioned to the BBC that the project team refers to hedgehogs with trackers as "digital hedgehogs".

To accurately identify these small animals and various objects in imagery, Tessera was trained with an enormous amount of data—approximately 20 petabytes or the equivalent of 10 billion regular digital photographs.

After reaching the limits of the university's computing capacity, extra processors were added under desks to maintain progress.

A recent collaboration with American tech companies AMD and Vultr has also provided additional computing resources.

Tessera's comprehensive satellite imagery, shown here capturing six global regions, is instrumental for mapping and analysing changes across diverse landscapes and terrains.

The extensive, open-source Tessera system is about more than just hedgehogs.

Cambridge University's Professor of Planetary Computing, Anil Madhavapeddy, indicated that over 100 research entities have utilised the tool.

This open-source system is also used for monitoring agricultural land, discerning which crops are cultivated in specific fields over time, and building a comprehensive picture of agriculture in the UK.

"Using satellite data can be very challenging due to its complexity and noise, such as needing to account for clouds and the difference between day and night," he said.

"Tessera simplifies this by condensing vast datasets into user-friendly maps of the UK, enabling us to ask precise questions about visible phenomena from space."