Experience how robots and humans can collaborate — linking imaging and AI (artificial intelligence) to plant science, transforming the future of our gardens and crops.
PhenAIx is our robot phenotyping system.
Phenotyping is the process of identifying observable characteristics in an organism which reflect its genetic composition. Phenotyping within Plant Science helps plant breeders select genetic varieties that exhibit desirable characteristics, such as high yield (for crops), beautiful colouring (for flowers) and/or resilience to disease.
PhenAIx demonstrates how robotic sensors and AI modelling can aid this process, by using specialist digital cameras to automatically capture both visible and non-visible plant properties and translate those into metrics used to assess plant quality.
Traditionally, people undertake this process, but there are limitations: people can only track visible properties, and people get bored with repetitive tasks and make mistakes. Imagine having to count the leaves in 100 plants in your breeding lab every week! A robot system like PhenAIx can do this quickly and accurately, without getting bored. Then the data collected by the robot system is shared with the human experts–who assess the metrics and ultimately make breeding and design decisions.
Watch a short video of PhenAIx in operation:
Plant Science
Our plants
We have carefully selected a diverse range of plants that demonstrates how the technology on display can be applied to species with different traits, aesthetic features or cropping requirements.
Ornamental plants
Agapanthus praecox subsp. orientalis [Zambezi] (‘Kek5006’)
Agapanthus praecox is native to the Kwa-Zulu Natal and Eastern Cape provinces of South Africa. This new variety, selected for its distinctive variegation, was celebrated amongst the shortlist of entries in the New Plant competition Chelsea Flower Show 2025. Kindly supplied by Pinnacle Plants International.
Fatsia japonica ‘Tsumugi-shibori’
Fatsia japonica, also known as Japanese aralia, is a species of flowering plant in the family Araliaceae, native to southern Japan and southern Korea. Raised and named ‘Tsumugi-shibori’ in Japan 1963 by Kawarada san, it has since become popular in gardens across the Western World under its alternative name, Fatsia japonica ’Spider’s Web’. Kindly supplied by Pinnacle Plants International.
Yucca aloifolia red experimental (yet to be named)
Yucca aloifolia or Spanish bayonet is a native of Mexico, Bermuda, and the Caribbean. It is noted for its sparsely branched trunk that produces iconic rosettes of sharply pointed leaves. Yet to be named, red experimental is an exciting new selection with bright red foliage that is in development at Pinnacle Plants International’s tissue culture laboratory. Kindly supplied by Pinnacle Plants International.
Food crops
The two food crops species on display have been grown in the University of Lincoln’s geothermal glasshouse, the first of its kind research facility in the UK. The facility uses geothermal ground source heating technology to provide heating from renewable energy.
This bushy tomato variety, typically selected for growing in pots or baskets, is used to represent the importance, significance and opportunities for UK tomato crop production. Future breeding programmes will continue to select varieties for flavour and their suitability for new growing systems and technology.
Fragaria x ananassa ‘Malling Centenary’
The strawberry plants shown in our exhibit are part of University of Lincoln’s glasshouse research. Glasshouse production enables an extended production season to support sustainable, consistent, home-grown supply by regulating temperature, light, humidity and carbon dioxide.
Plant Traits
The PhenAIx system automatically measures the following plant traits, which can help plant scientists, plant breeders, agronomists and gardeners track the features (e.g. appearance), behaviour (e.g. response to growing environment) and health (e.g. presence or absence of disease) of different plant varieties in the lab and the field.
Here’s the plant we start with:
Morphological Traits
Morphological, or architectural, traits refer to metrics that describe the size and shape of a plant. These can be used to track growth, as well as compare varieties.
Trait
Description
Example
Leaf Count
A trained AI model analyses the image and detects and counts individual leaves on the plant. This is an application of Machine Learning and Computer Vision, where a computer model has been created through training with many data examples to recognize and quantify leaves in various conditions.
3D Point-Cloud Model
A point cloud is a 3-dimensional (3D) model marking the extents of a volumetric shape. The depth camera helps construct the point cloud by measuring the distance from its lens to the outer edges of the plant, gathering a collection of points that together comprise an approximation of the plant’s shape. In the live PhenAIx display, the points are collected over time and the 3D point-cloud model fills in more detail as the system runs.
Spectral Traits
Spectral traits, or vegetative indexes, refer to metrics that describe properties illuminated by light wavelengths outside of the human visible range, known as hyperspectral (or multi-spectral, which is a subset of hyperspectral wavelengths).
Infrared (IR)
IR images are quick to capture in our PhenAIx setup. Infrared light can be used for thermal imaging and may reveal information about plant health.
Near-Infrared (NIR)
NIR light is invisible to the human eye but strongly reflected by healthy green leaves. NIR is a key input for calculating vegetation indices such as NDVI.
Red Edge
The red-edge band bridges red and near-infrared wavelengths. It is highly sensitive to chlorophyll concentration and provides early warning of plant stress before it is visible to the naked eye.
Normalised Difference Vegetation Index (NDVI)
NDVI maps vegetation health by comparing red and near-infrared reflectance. Green indicates thriving plants; red areas suggest stress or bare soil. NDVI can also be used to measure levels of variegation in leaves.
Robotics and Artificial Intelligence (AI)
Our Technology
PhenAIx is our robotic phenotyping system, designed as a versatile outreach exhibit that can bring cutting-edge agricultural robotics directly to schools and community events across the UK. This unique, self-contained robot-in-a-box system features a compact robotic arm operating safely within a confined space. PhenAIx is equipped with multiple advanced sensors including multi-spectral and depth sensors. The system is designed to actively scan live plants in real-time and detect traits that are used by plant breeders, agronomists and gardeners in plant variety selection.
The exhibit showcases how modern AI and robotics are revolutionising sustainable farming through automated plant phenotyping–the measurement and analysis of plant characteristics that indicate health, growth potential and environmental resilience. Crucially, PhenAIx demonstrates how robotics and AI complement rather than replace human expertise. PhenAIx utilises multi-spectral imaging, depth sensing and AI-driven computational analysis to provide insights beyond un-aided human sensory capabilities. Our novel system demonstrates techniques developed, applied and refined through the University of Lincoln’s UKRI-funded research programmes.
Watch as our robotic arm autonomously scans plants, measures key morphological traits (such as leaf count) and spectral traits (such as NDVI, Normalised Difference Vegetation Index).
Design
Our unique robot-in-a-box design features a robotic arm and two specialised camera sensors (individual components are described below), moving in tandem with a rotating pedestal. When a plant is placed on the pedestal, the camera sensors capture features of the plant from many different angles, above and around all sides as the plant rotates.
The sensor data is fed into a Computer Vision (CV) model, which analyses the images captured and produces metrics that describe the plant size and shape, health properties and other features–the set of observable traits that comprise the plant’s phenotype.
Computer Vision is essentially comprised of two components: Machine Vision, which entails mathematical analysis of digital images, combined with Machine Learning (ML), a form of AI in which computational models are learned, or generalised over time, based on a set of training examples.
An iterative process is undertaken whereby the machine learner builds up a theory about the contents of its training set.
The aim is for the learner to acquire a general model that can be applied to any relevant data set, comparing key features in the training set to features in the model.
One example included here is the leaf count trait.
The PhenAIx system learns how to recognise leaves of different plants, as captured by its cameras.
Human engineers help train the model, by providing a set of correct examples that the machine learner uses for training.
Ideally, the learned model is generalisable: for example, if trained on strawberry images (i.e. to recognise leaves of strawberry plants), can the model also be applied to other plants, such as tomatoes? Answering this question in many different contexts is the subject of much research–including research conducted by our team at the University of Lincoln.
Universal Robot’s UR3e arm is light-weight and compact, offering 6 degrees of freedom (6-DOF) via joints at the base, shoulder, elbow and three wrists. Within PhenAIx, the UR3’s end-effector (end-point) is fitted with sensors (described below) and the arm can rotate above and around the plant, to capture data spanning top, side and underside views.
The MicaSense RedEdge-MX camera provides multi-spectral imaging using Blue, Green, Red, Red edge and Near infrared (NIR) wavelengths.
People
Our Team
Our PhenAIx @ Chelsea Flower Show team includes Robotics and AI experts, Plant scientists and a dedicated support team, spanning a broad range of STEM careers and career stages, including PhD students, Post-Doctoral Research Associates, Professors and Professional Services staff.
We are based within the Lincoln Institute of Agri-food Technology (LIAT) and the Lincoln Centre for Autonomous Systems (L-CAS) at the University of Lincoln.
Meet our team members below!
Plant science and horticulture experts
Dr James Wagstaffe
James is Director of Teaching and Learning at LIAT, working across ornamental and food crops. Passionate about inspiring people to pursue careers in horticulture, he champions the broader benefits plants bring to communities and society. He also serves as a Trustee of the Chartered Institute of Horticulture, leading its Education Committee.
Dr Yoon Cho
Yoon is a Post-Doctoral Research Associate in Plant Physiology who has led the development of the robot-assisted plant phenotyping pipeline. Her research centres on evaluating plant physiological traits using data collected by phenotyping robots. Her expertise incorporates high-throughput data analysis and machine learning of 3D and spectral data.
Krystian Lukasik
Krystian is LIAT’s Technical Resource, specializing in management of hydroponic crops in controlled environments. At the University of Lincoln, he works with students to develop their practical skills and collaborates with teams conducting research on viticulture, polytunnel crops and the university’s (the UK’s first!) research-focussed geothermal glasshouse.
Rob is a Senior Mechatronics Engineer who has overseen the design and construction of the PhenAIx system, from the box containing the arm to the hardware connectivity.
Dr Ollie Hardy
Ollie is a Post-Doctoral Research Associate in Intelligent Systems and has contributed to the hardware implementation and testing of the system. Ollie is interested in field robotics and novel data collection.
James is a PhD student within the EPSRC Centre for Doctoral Training in Agri-Food Robotics: AgriFoRwArds and a Research Associate in Fleet Coordination. His research focuses on building systems to organise routes and manage resources shared between autonomous robots in new agricultural environments. He has provided technical support for PhenAIx.
Simon (Pearson) is Professor of Agri-Food Technology and Head of LIAT’s School at the University of Lincoln, Founding Director of LIAT and MBE recipient for Agricultural Innovation. Leading £110M+ in agri-tech projects, his work spans AI, robotics, and food systems, influencing national policy, industry, and global forums on the future of agri-food.
Professional Services
Lyn Heaton
Lyn is LIAT’s School Manager, ensuring smooth day-to-day operations while supporting students, academics, and professional staff. She coordinates external visits, events, and senior-level liaison. Known for calm professionalism and practical problem-solving, she also served as Project Manager for the exhibit, keeping the team organised and on track.
Heather Smith
Heather is LIAT’s business development and stakeholder engagement lead, a professional who is passionate about translating research into accessible content for external audiences. Focused on promoting STEM and agri-tech innovation, they bridge the gap between cutting-edge research and real-world impact, driving awareness and collaboration across diverse sectors.
Minnie Haigh
Minnie is the LIAT Administrator, supporting students from enrolment to completion and providing high-quality administrative services to staff and academics. She also coordinates events and short courses. With a background in graphic design, she has contributed creatively to projects, including materials for this exhibit.
Dr Paula Eves
Paula is a senior leader overseeing operations, infrastructure, health and safety, and capital developments at Riseholme LIAT. She drives research development, income generation, and interdisciplinary collaboration, while advancing agri-tech innovation through industry partnerships and research translation. She also leads health and safety for the exhibit.
Exhibit Information
Sustainability
In designing and building our exhibit, sustainability has been at the heart of our discussions and project decisions. We expect that many of the constituent parts of our exhibit will be used in future activities, for example, we anticipate taking our robot to schools and outreach events to promote the importance and opportunities for STEM subjects in the horticulture and agriculture industries. We considered sustainability throughout or exhibit, for example:
Planter
The wooden planter has been kindly and carefully constructed by Mark Close using recycled scaffolding boards. Using recycled wood in garden construction is a good idea because it reduces demand for newly harvested timber, helping to conserve forests and lower the environmental footprint of the project.
Glasshouse structure
Our greenhouse structure has been kindly constructed by CambridgeHOK. CambridgeHOK built the University of Lincoln geothermal greenhouse and much of the material used here has been recovered from other construction projects.
Leaflets and Plant Information Cards
The leaflets and plant information cards are printed on FSC recycled paper stock.
Wood finishes
The finish on our worktop is beeswax and the finish on our planter is linseed oil. Both are naturally occurring substances.
Sponsors
We are very grateful to our sponsors, without whom PhenAIx and our Chelsea Flower Show exhibit would not have been possible:
Contact
For further information, please contact LIAT Info.
