As soil health deteriorates and labour costs rise, smart farming is increasingly looking to technology and AI to boost yields. Over the last decade, the technology has evolved from satellite navigation for autonomous tractors to AI and data science to the latest satellite communications. At the same time there is a drive to use agritech to reduce the amount of pesticides, fertiliser and water being used to reduce contamination and carbon dioxide emissions
However with farming particularly fragmented across Europe, the focus has shifted to the data that is gathered from sensors so that it can be used by AI models to boost production is many different ways.
Slugs
One of these is a UK project is that is using hyperspectral cameras to generate risk-prediction maps for AI monitoring that helps farmers battle slugs.
SLIMERS – Strategies Leading to Improved Management and Enhanced Resilience against Slugs – is a three-year £2.6m research programme involving more than 100 farms and seven partners, including Harper Adams University, the UK Agri-Tech Centre, the John Innes Centre, Fotenix and Farmscan AG.
One of the slugs in the trial Courtesy of BOFIN
The project is in its last year developing cost-effective forecasting and precision treatment tools, including Al-based autonomous slug monitoring and biological control and exploring ‘slug resistant’ wheat varieties.
The AI model has been created with data from farmers’ slug monitoring activities over the previous two years of the project, combined with extensive soil mapping and testing, the model predicts areas in fields with a high likelihood of containing slugs.
The next step is to put the model to the test, using it for selective applications of slug pellets rather than blanket application. The data collected will also be used to further develop the model.
“We’ve known for some time that slugs gather in patches, but prior to SLIMERS we didn’t understand fully the specific factors that cause this and how the patches can be reliably located,” said Professor of Invertebrate Biology and Pest Management Keith Walters, who leads the work at Harper Adams.
“Thanks to the data collected we now have that understanding and are using our predictive model to produce detailed risk maps for their fields. In 2025-6 we are asking them to treat only the predicted slug hotspots to fine-tune the models and bring the vision of precision pest management closer to reality.”
“Our role is to build AI-powered slug detection, right there in the field. But first, we’ve got to train the AI, and that means putting slugs in the crosshairs,” said Charles Veys , CEO of Fotenix, which develops an AI-enabled hyperspectral camera. This has been used to pinpoint the exact spectral signature of the unwelcome visitors. As slugs don’t tend to surface until after dark this has meant heading out in the small hours to collect the training imagery data.
The team at Farmscan AG in Manchester are working on a system with the smallest spray width possible which will be added to an existing autonomous farm vehicle to kill the slugs automatically.
“We are aiming for 25cm or less, which would mean four nozzles per metre,” said Callum Chalmers, director at Farmscan. “We are running the first trials at the end of 2025, then field trials will be in full swing in early 2026.”
This will be used on a system developed by the Small Robot Company.
“Robotic slug control is a natural next step for Small Robot Company. Precision robotics has tremendous potential for agriculture across the board. Slugs are a terrible bane for farmers. They can decimate emerging crops and have a significant impact on yield. But treatment is problematic. Both water and wildlife have been impacted by chemical methods. Legislation is looming — farmers need an affordable alternative,” said Ben Scott-Robinson, CEO and co-founder of the Small Robot Company.
“What’s remarkable is that we will soon have a service that we can roll out as a tool that all UK farmers will come to rely on to reduce their reliance and expenditure on pellets to control arable farming’s biggest pest. With increasing pressure on chemical control, finding sustainable and environmental solutions has never been more important,” said Tom Allen-Stevens, managing director of the British On-Farm Innovation Network which leads the project.
Goats
Researchers in the UK are testing out wearable sensor technology for improving the health, welfare, and management of dairy goats.
The project is led by Dr Holly Vickery from Harper Adams University with Dr Gemma Charlton from the Animal Behaviour and Welfare Research group alongside Dr Zoe Barker of Reading University and SmartBell in Cambridge.
Dr Holly Vickery with one of the goats in the trial Courtesy of Harper Adams University
Despite their significance, goats remain underrepresented in technology and welfare research. This study addresses that gap by assessing whether wearable sensors—widely used in dairy cattle—can reliably monitor key goat behaviours like lying, feeding, and rumination.
“Monitoring animals as individuals, rather than as groups, is essential to ensure high welfare standards,” said Vickery. “Wearable sensor technologies offer an innovative way to track the wellbeing of each animal without increasing labour demands. For goats, this could mean early detection of health issues, better responses to environmental stressors such as heat, and more efficient, precise management strategies.”
A trial sees 40 SmartBell ear tags fitted to goats on a cheese-making farm in Somerset. These integrate an accelerometer to see whether an animal is lying down or unsteady on its feet, as well as ear and ambient temperature sensors with a wireless link to a gateway that delivers the data into a cloud-based system.
This system implements a fully-automated solution for livestock management, including distress alerts, herd analysis and customised farm management reports. Using advanced analytics on data gathered from animal wearables, it also cross references regional data.
The study is looking at the current knowledge of wearable technology within the dairy goat industry and identifying perceived barriers, to validate the effectiveness of ear tag sensors in monitoring goat behaviours on a commercial farm, and to co-develop a roadmap for future technology implementation and research priorities with industry.
Precision agriculture
The 5GRIT project in the UK has been testing out drones remote surveying of grazing livestock, primarily sheep, on a 600Ha farm outside of Slaggyford, Northumberland. Over a 5G network, the in-field transmission of survey data to a remote cloud server was completed around three to four times faster when compared with existing 4G data speeds.
The project also developed an AI livestock detection algorithm to survey and assess livestock and achieved 86% accuracy. Indications suggest that this accuracy value could be further improved through additional data-set training and optimisation for wider detection targets, allowing for further livestock analysis. Additional cost efficiencies could be enabled through regular remote aerial surveillance, facilitating early identification of livestock health and enabling farm security benefits.
Meanwhile farmers in Ireland have been using the Internet of Things to improve the efficiency of conception. Startup MooCall produces IoT sensors for cows where a collar worn by a stock or a teaser, also known as vasectomised bull and ear tags are attached to all cows and heifers.
The collar then uses the proximity of the cow or bull, mounting behaviour and activity levels to determine when a cow or heifer is in heat to a high degree of accuracy. It can also identify likely due dates and monitor a cow’s fertility over time. Since using the technology, Ciarán Lenehan, a suckler farmer from Co. Meath, has seen conception rates on his farm increase from 75-85% to over 90%.
Harper Adams University is also conducting research into the application of machine learning. To predict pen fouling, diarrhoea and tail biting in commercially farmed indoor pigs, its RoboChick project has developed a small autonomous vehicle for monitoring. This sits at approximately bird head height and is capable of collecting high-resolution data such as temperature, humidity, carbon dioxide level, air flow velocity and light spectrum
Water scarcity
Water is another scarce resource where sensors and data are vital. UK-based farm equipment giant CNH has been worked with Italian agritech company xFarm on a smart farming pilot project in Uzbekistan to show how precision technology and data from its machines can help farmers increase productivity and save water.
The Drops of the Future initiative is the first time the Organization for Security and Co-operation in Europe (OSCE) has worked with private companies on the topic of water, agri-food production and energy.
“Today, agriculture uses over 70% of the world’s freshwater. We believe that partnerships, such as our collaboration with OSCE, are key to unlocking the full potential of solutions that benefit both farmers and the environment,” said Daniela Ropolo, Head of CNH EMEA’s Sustainable Initiatives.
“By combining expertise, resources, and technology, we empower farming communities to use water more efficiently, boost food production, and reduce environmental impact. Partnering with OSCE has been a great opportunity to advocate for lasting, environmentally conscious change in agriculture.”
The project involves machine telematics and the use of FieldOps, the all-in-one farm management platform that enables farmers to connect, view and manage operations. In total, the area covers 6,000 hectares – equivalent to 7,000 football pitches – where a variety of crops including, cotton, wheat and rice are being grown.
The machines in the field include 48 Case IH Puma tractors, two Case IH Patriot Sprayers, 15 Row Cotton Pickers and two New Holland TC combine harvesters, all produced by CNH.
“OSCE’s Drops of the Future project focuses on equipping farmers with the tools and knowledge to adopt precision farming techniques that improve water efficiency and reduce environmental impact in Central Asia,” said Gianluca Feligini, Head of Precision Technology at CNH. “Through workshops we are able to share our knowledge and experience of the technology in our machines that can directly benefit farmers.”
Adopting and embedding AI capabilities is a key new strategy for CNH announced in May to integrate the technology into its equipment. This includes agronomic sensors, smart implements, advanced automation, autonomous features, satellite connectivity, agronomic insights, and machine data synchronization via its FieldOps digital platform.
FieldOps is designed to enable farmers to view and monitor all their CNH machines in one place, monitoring every machine in real-time, remotely viewing in-cab displays to deliver better feedback to their operators and giving visibility of layers of agronomic data — all in one platform.
To make data management even easier for farmers, FieldOps has over 40 API (application programming interface) connections available globally, creating seamless integrations with third-party digital platforms.
Starlink satellite connectivity
A key part of smart farming is connectivity, either via telematics modems or more recently through Starlink satellite data receivers. This provides real-time data and remote support to farmers without the hassle of managing or renewing a subscription.
This is a completely new approach to connectivity that makes data accessible by simply making it part of the machine, says CNH. As a baseline feature, it expands the reach and use of connected features, helping farmers improve their productivity and use of agronomic data using the data from their machines, fleets and fields.
The FieldXplorer platform uses AI to transform drone images into a field map that distinguishes between weeds and crop. The deal with StarLink in May 2025 allows that data to be exported nearly instantaneously, for example to create a prescription spraying map for the machine. This enables farmers to apply pesticides sooner, controlling weeds earlier and so using less chemicals, which ultimately helps improve crop yields.
“With these advancements in our open digital ecosystem, we add new milestones in our ongoing global innovation journey to make everyday operations more efficient and productive for farmers. These projects are designed 100% around our customers and dealers, to provide data and help them increase productivity and reduce costs,” said Carlo Lambro, brand president of New Holland, part of CNH.
The real-time monitoring data transfers approximately three to five seconds following each machine action and location change. Real-time monitoring features are currently available on the T7 Longwheel Base, T7 Heavy- Duty, T8, and T9 tractor models but will be phased in across other platforms, including combines and sprayers, in future updates
CNH has been developing the NHDrive and the more recent Raven OMNiDRIVE autonomous control systems that can be fitted to existing tractors, but also self driving tractors without a cab. At the same time its US competitor John Deere has been buying up autonomous control technology by buying Bear Flag Technologies, Blue River and fast charger specialist Kreisel Electric in Austria.
“With the integration of Raven’s driverless agricultural technology, the platform has added driverless capabilities in harvest applications, which has tremendous benefits in today’s labour-constrained market,” said Eric Weaver, New Holland Precision Technology Global Director New Holland.
“Driverless technology is now a reality and fully integrated in New Holland’s products thanks to Raven technology. The state of regulations on autonomy today is diverse around the globe and we cannot predict how regulations are going to evolve, so we are preparing for different scenarios: an approach that will allow us to be very agile as regulation allow market introduction. In essence, the path to autonomy is a journey, and this new integration is the demonstration of the efforts we put into continuous innovation.”
While CNH showed its cabless tractor was shown a decade ago, now the focus for smart farming has shifted to the data that is generated by all the technology to feed the control systems for the tractors.
AI data
The current agrifood data landscape consists of data scattered in different places and forms, subpar dataset search capabilities, and data unfit for AI tasks. This falls short of user needs and projects such as Farmtopia and STELAR as aiming to tackle the issue.
Farmtopia is developing nine sustainable innovation pilot (SIP) projects across Europe to boost production of carrots in Italy, diary cows in Romania, hemp in Lithuania and mushrooms in Hungary, Germany and Ireland.
Meanwhile the STELAR project will design and develop a Knowledge Lake Management System (KLMS) for turning raw data lakes into knowledge lakes. This combination of a platform and its tools will also hold the purpose of simple and intelligent data discovery, AI-ready data, and semantic interoperability in smart agriculture and food safety applications. Through all of this, STELAR will embrace the FAIR (Findable, Accessible, Interoperable, Reusable) approach to data.
One of the pilot projects in STELAR is integrating satellite, hyperspectral, meteorological and synthetic data for early crop growth predictions, while the other is Integrating different types of sensor technology data from the field.
New technologies often face challenges in varied environments; for instance, camera-based systems may underperform due to lighting, soil, or weather differences. Setting up these machines can also be difficult as conditions change by location.
Improving data discovery, linking, and annotation for agrifood datasets in smart farming to handle issues of volume, diversity, and quality aims to tackle the scattered and inconsistent data. This is a key step for making the most of the data and allowing the practical use of autonomous systems across all kinds of farms.
