TomatoGuard uses AI-powered digital nose sensor to detect early signs of crop stress
A new research and development project aims to introduce a cutting-edge AI-powered monitoring system to facilitate early detection and response to crop stress in protected crop production.
Backed by Defra and Innovate UK, the TomatoGuard: Advanced AI-Driven Pest and Stress Detection for Sustainable Tomato Cultivation project combines expertise of The UK Agri-Tech Centre (ATC), Altered Carbon (AC), Fargro Ltd and grower APS Produce, to develop the machine learning required to detect specific tomato volatiles indicative of stress.
The technology at the heart of TomatoGuard consists of AC’s pioneering AI-assisted Digital Nose sensor system, which emulates human senses to identify potential horticultural issues around the plant and in the soil. It uses a graphene-based sensor array to detect specific gaseous elements and vapours, serving as an early warning system to identify environmental changes that indicate the early stages of impending crop stress.
The technology is based on AC’s K9sense chip, a new type of complex chemical sensor, which works within an AI framework to detect unique volatile organic compound (VOC) patterns released by plants under stress.
“Using this smart sensor to spatially detect crop stress early will allow effective measures to be deployed leading to reduced pesticide use, crop loss and more sustainable and economic crop production,” ATC said.
AC also aims to lower the carbon footprint of running a protected cropping system by detecting plant and soil stress signals for fine-tuning the environment’s heating, humidity and ventilation.
To have maximum impact, the project partners have chosen to work on one specific issue affecting tomatoes, and general anomaly detection. The specific detection will target Red Spider Mites (RSM) – one of the most prolific pests in tomato cultivation.
Although biological treatments exist, the organisers of the project said their effectiveness will be greatly increased by early identification. Many effective chemical options have been withdrawn from the market in recent years due to environmental, food safety and commercial reasons.
With a combination of supervised (RSM) and unsupervised (anomaly detection) learning, TomatoGuard uses a Machine Learning Model, which promises to significantly innovate agriculture in the UK.
Initially, TomatoGuard will focus on tomato cultivation, a choice driven by the crop’s economic significance and unique agronomic challenges. However, the technology is designed to be versatile and scalable, with potential applications in other protected crop environments including vertical farming.
With an extensive testing strategy that includes laboratory-controlled environments, glasshouse trials and commercial growing conditions, AC aims to ensure the robustness and reliability of the TomatoGuard system.
“TomatoGuard transcends traditional technology-based solutions, bridging the gap between agronomic, commercial expertise and state-of-the-art sensor technology,” AC said.
The consortium envisions this project as a tool that empowers growers, facilitating a shift towards a more sustainable, efficient system for crop production.
Dr Andy Evans, innovation lead for crop health at the UK Agri-Tech Centre, said: “The outcomes from this project will lead to the detection of issues arising within tomato crops at a very early pre-symptomatic stage, allowing interventions to be deployed in the areas where issues are arising.
“From a grower perspective this would cut down on costs due to unnecessary pesticide use by targeting the areas that need treatments, optimising the nutritional requirements of the crop and maintaining a high quality and yielding product. The system has the potential to be applicable in a wide range of protected cropping systems.”
Detelina Vassileva, chief operating officer at Altered Carbon, said: “Our mission is to improve life on Earth by digitising the sense of smell. Food loss is a huge issue affecting 30 per cent of all that we produce worldwide, and it is during crop production that we can have the biggest impact.
“We are hoping to expand on this one pest, and one crop, into many more and have real measurable impact on the quality and quantity of the food that is produced worldwide.
Lori-Leah Griffiths, technical and regulatory specialist at Fargro, said: “Fargro’s role in the TomatoGuard project is centred on providing expert agronomic advice to guide the development of this innovative technology and independently monitor commercial trials to ensure its practical application for growers.
“By combining our deep horticultural expertise with advanced sensor technology, we aim to deliver solutions that empower growers to act early, target treatments effectively, and ultimately improve both environmental sustainability and economic outcomes.”
Brian Moralee, growing manager at APS Produce, said: “We believe that this new technology should increase our effectiveness of identifying pests and crop stresses at a much earlier stage, which will allow earlier interventions with cultural and biological controls. We hope in turn, that this will reduce our ever-increasing costs and help us meet our goals of more sustainable production”.
Trials have already started within glasshouse and controlled chambers at Rothamsted Research, undertaken by ATC, with further trials in a larger glasshouse at Stockbridge Technology Centre scheduled in Spring 2025.
During the project, APS Produce will also utilise the sensors in a commercial glasshouse setting in order to provide feedback on the use of the technology within the commercial growing sector.