Using AI to create a knowledge bank of plants that benefit the environment
The RHS will use AI (artificial intelligence) to help build a knowledge bank of cultivated plants for specific uses, such as pollination, pollution capture and water management, as it launches a new five-year programme of work to commence in 2025
Plants for Purpose will see the RHS work in collaboration with the University of Nottingham to develop a deep learning tool that will identify characteristics among the more than 400,000 different plant cultivars found in the UK that RHS and wider industry research has shown to be beneficial. (Deep learning is a method in AI that teaches computers to process data in a way that is inspired by the human brain.)
For example, RHS research has previously revealed that rough surfaced leaves are most adept at capturing particulate pollution. Machine learning techniques that reliably quantify texture can be used to identify specimens from the RHS Herbarium – a vast collection of dried plant specimens – with hairs and scales. These plants’ trait matches will be cross checked by botanists and horticultural scientists and definitive guides will be published over the course of the next five years.
Dr Michael Pound, Associate Professor, School of Computer Science, University of Nottingham, said: “It’s tremendously exciting to be part of a project that will deliver real impact for UK biodiversity. Deep learning excels when we have lots of data, and this is exactly what the RHS Herbarium offers. The trained networks will study hundreds of thousands of images to draw out the key features that distinguish plants. These can then be embedded into tools that allow RHS scientists to identify potentially important varieties in seconds rather than months.”
AI to feature at RHS Chelsea 2025



