'VespAI: Applying Deep Learning to the detection of Invasive Hornets' with Thomas O' Shea-Wheller, hosted by the CBKA

About the talk

The invasive hornet Vespa velutina nigrithorax is a rapidly proliferating threat to biodiversity and apiculture in Europe, East Asia, North America, and Oceania. To date, authorities have struggled to contain the hornets, as colonies must be detected and destroyed early in the invasion curve if establishment is to be prevented. The current UK response relies upon visual alerts by the public, however ~99.9% of submitted reports are in fact misidentified native species. With ongoing annual incursions, there is a pressing need to develop improved monitoring technologies within a limited time-frame.

In this talk,  Thomas will outline VespAI, an automated system for the rapid detection and behavioural quantification of V. velutina, V. crabro, and V. orientalis. VespAI uses an AI approach to detect hornets in real-time, and send image alerts via a compact remote processor.

Crucially, unlike traditional trapping methods, this eliminates any potential harm to native insects, while ensuring that detected hornets can be rapidly tracked back to the nest.

Thomas will discuss the development, performance, and future deployment of the system, and highlight its potential to enhance the surveillance and control of invasive hornets.

About Thomas

Thomas is interested in the complex interactions that govern collective behaviour, ecology, and self-organisation within social insects. As a Research Fellow based at the University of Exeter, he works with honey bees, bumble bees, and a variety of other social insects to improve colony health and resilience. His current research includes projects pertaining to honey bee epidemiology and Varroa resistance, bumble bee ecology, and the detection of invasive hornets using artificial intelligence.