Contrastive Learning of Plant Traits for Ecosystem Services
This University of Nottingham project, supported by the Royal Horticultural Society, explores the use of machine learning to uncover and explore plant traits within the RHS Herbarium. Plants play an important role in our ecosystem, providing services like air purification, water retention, and urban cooling. By training algorithms on digitised herbarium specimens and live plant images, the project aims to create software that can identify key morphological features linked to these environmental benefits. This tool will empower researchers at the RHS to search for and identify new species of plants that may have positive impacts on our environment, and work with gardeners across the UK to propagate these species in gardens. The project is part of the RHS Plants for Purpose initiative, the research will inform new plant recommendations, broaden the palette of sustainable garden species in the UK, and support biodiversity.
CVL people - Mike Pound