AI guards against invasive hornets

Insect diversity and abundance are in decline in the UK and threatened further by invasive species. Invasive species colonise areas they are not native to, disrupting the usual ecosystem balance. This can decimate other species, including those that carry out important roles like pollination.

The invasive yellow-legged hornet (also sometimes known as the Asian hornet), Vespa velutina, was first sighted in France in 2004. Twenty years on, the species has made approximately 500,000 nests there, predating on native insects and putting pollination, as well as the honey industry, at risk. Whole honeybee hives can be lost.

In 2016, the first yellow-legged hornet was sighted in the UK. To control the invasion, detection of Vespa velutina is key but currently relies on public records. Only 0.01% of those sightings are correctly identified.

Professor Juliet Osborne at the University of Exeter is leading a team of researchers to develop a Vespa velutina detection system called VespAI. VespAI uses a camera, paired with an AI-based image recognition method, to identify the hornet as it visits a baited monitoring station. Once detected, the system sends an alert to users, enabling control measures to be taken.

In 2024, VespAI was tested on the island of Jersey, where Vespa velutina visited regularly, and also in Kent and Hampshire, where sightings are rare. VespAI proved to be highly accurate, distinguishing Vespa velutina from other insects with over 99% accuracy. The technology (which monitors insects continuously) offers a promising means of detecting the species at the frontline of the invasion, supporting the government’s actions to tackle this invasive species.

This research was supported by BBSRC responsive mode and Follow-on Fund funding.