AI Assists in Detecting Invasive Asian Hornets

New research showcases the use of Artificial Intelligence in detecting invasive Asian hornets, highlighting a potential threat. Exeter researchers developed VespAI, an automated system attracting hornets to a monitoring station, capturing standardized images with an overhead camera.

VespAI, from University of Exeter researchers, accurately detects Asian hornets, facilitating rapid responses. These hornets, also called yellow-legged hornets, have infiltrated mainland Europe, East Asia, and recently, regions like Georgia and South Carolina in the US. Given the UK’s vulnerable position and annual incursions, improved monitoring systems are imperative.

Designing VespAI as an Inclusive Solution

Dr. O’Shea-Wheller stressed creating a versatile system for use by governments and beekeepers. The prototype, VespAI, demonstrated promising results as a robust early warning system for detecting Asian hornet incursions into new areas.

VespAI, equipped with a compact processor, remains inactive until it detects an insect of hornet size. Upon detection, its AI algorithm distinguishes between Asian hornets (Vespa velutina) and native European hornets (Vespa crabro) in the captured image. Upon recognizing an Asian hornet, the system notifies the user with an image for verification.

Currently, the UK’s response strategy relies on individuals spotting, identifying, and reporting Asian hornet sightings, which has inherent limitations.

Addressing Misidentification

Dr. Peter Kennedy, the individual behind the system’s concept, lamented, Unfortunately, the majority of reports received contain misidentifications of native species. As a result, responsible agencies are tasked with manually confirming thousands of images each year.

Our system is crafted to provide vigilant, accurate, and automated surveillance to tackle this challenge.

Dr. O’Shea-Wheller noted that in some European areas, hornet detection relies on traps, causing unintended harm to native insects and offering limited control over Asian hornet populations. In contrast, VespAI minimizes environmental impact by sparing non-target insects. Moreover, it captures live hornets, facilitating their tracing back to nests, the most effective eradication strategy.

VespAI underwent rigorous testing on Jersey Island, where frequent Asian hornet incursions occur due to its proximity to France. Despite encountering diverse insect species, including Asian and European hornets, VespAI’s detection algorithm effectively distinguished between them, even amidst significant numbers.

“The system’s precision is its main advantage; it doesn’t misidentify other species or overlook any visiting Asian hornets,” Dr. O’Shea-Wheller emphasized.

Strengthening Exclusion Measures against Asian Hornets

However, biologists and data scientists from the University of Exeter, along with partners from Defra, the National Bee Unit, the British Beekeepers Association, and Vita Bee Health, are deploying additional prototypes. This initiative aims to bolster exclusion efforts amidst the surge in Asian hornet sightings in the UK in 2023.

Alistair Christie, Senior Scientific Officer for Invasive Species in Jersey and part of the testing team, expressed, “The proposed device may prove a powerful tool in the early determination of the presence of Asian hornets in an area, and thereby fills an important gap.”

To conclude, the research paper, titled “VespAI: a deep learning-based system for the detection of invasive hornets,” was published in the journal Communications Biology.

Read the original article on: Phys Org

Read more: Beekeepers can Now Inspect Hives Without Disturbance