AI Takes on Hidden Crop Threats in UK Onion and Brassica Fields

A new grower-led project is turning to artificial intelligence and smart trapping technology to help British onion and brassica producers detect damaging pests before they take a costly bite out of yields.

TRACER-Pest will combine field-based cameras, machine learning and real-time data analysis to identify and track bean seed fly and swede midge — two pests capable of causing significant commercial losses but notoriously difficult to monitor using conventional methods.

The two-year feasibility project has received £330,058 through Defra’s Farming Innovation Programme, delivered in partnership with Innovate UK. It is being led by P.E. Simmons & Son, working alongside P.G. Rix Farms, Farmacy Ltd, the University of Warwick, B-hive Innovations Ltd and the British Growers Association.

Together, the partners bring commercial growing operations, agronomy, pest science, engineering and artificial intelligence into a single project designed around practical conditions in the field.

A race against the pest lifecycle

Bean seed fly larvae can damage seeds and young shoots, causing uneven crop establishment and potentially reducing marketable yields. Swede midge presents a different but equally difficult challenge, with treatments needing to be applied at a precise stage in the insect’s lifecycle.

Existing monitoring systems, including pheromone traps and manual inspections, can make it difficult to determine the best moment for action. They also require trained personnel to inspect traps, distinguish target pests and interpret changing populations across multiple sites.

TRACER-Pest aims to automate much of that process.

AI-enabled cameras attached to traps will capture images in the field, with machine-learning models trained to detect and classify the relevant insects. The resulting data should allow growers and agronomists to follow pest pressure in near real time and respond within critical treatment windows.

Commercial trials will take place across onion and brassica production sites. The technology will be compared with existing monitoring practices while researchers refine its image recognition, classification and reporting capabilities.

From routine checking to targeted action

If successful, the system could reduce dependence on manual scouting while providing more consistent information across growing sites.

The potential benefits extend beyond labour savings. Earlier detection could help growers protect yields, avoid unnecessary treatments and apply authorised crop-protection products more precisely.

That is becoming increasingly important as horticultural businesses contend with narrowing plant-protection options, rising input costs, workforce pressures and more unpredictable pest behaviour under changing climatic conditions.

Rather than replacing the experience of growers and agronomists, the technology is intended to give them faster and more objective evidence on which to base their decisions.

The project also reflects a wider shift in agricultural technology: away from collecting data simply because it is possible and towards providing growers with timely, commercially useful answers.

Findings from TRACER-Pest will be shared with growers and agronomists across the sector. Its longer-term value will depend upon whether the project can deliver reliable identification under variable field conditions — and do so at a cost that makes commercial adoption viable.

For Britain’s onion and brassica growers, however, the promise is compelling: pests identified sooner, interventions timed more accurately and fewer crops lost to threats that are too often discovered after the damage has already been done.