Pl@ntNet

Pl@ntNet
Name	Pl@ntNet
Developer	Tela Botanica; Institut National de la Recherche Agronomique
Released	2013
Operating system	Android; iOS; Web
Genre	Biodiversity; Citizen science; Image recognition
License	Proprietary; data under multiple licenses

Pl@ntNet

Pl@ntNet is a citizen science platform and mobile application for plant identification that combines image recognition, taxonomic databases, and community curation to document vascular plants, bryophytes, fungi, and algae. It integrates machine learning models, crowdsourced observations, and institutional floristic references to support research by botanists, ecologists, conservationists, and educators. The project operates through international collaborations spanning research institutes, universities, herbaria, museums, and nongovernmental organizations.

Overview

Pl@ntNet operates at the intersection of computational botany, biodiversity informatics, and citizen science, linking smartphone photography to taxonomic knowledge from institutions like the Muséum national d'Histoire naturelle, Royal Botanic Gardens, Kew, and Missouri Botanical Garden while leveraging machine learning research groups at INRIA, CNRS, and University of Cambridge. The platform aggregates observations into datasets comparable to those managed by GBIF, iNaturalist, and the Global Genome Biodiversity Network, and interoperates with standards from the Catalogue of Life, ITIS, and Tropicos. Users submit images that are processed by convolutional neural networks developed in collaboration with research groups at Carnegie Mellon University, École Polytechnique, and University of California, Berkeley, producing candidate identifications linked to taxonomic authorities such as APG, Flora of North America, Flora Europaea, and Flora do Brasil.

History

The project originated from research initiatives at Tela Botanica and agricultural research institutions including INRA (Institut National de la Recherche Agronomique) with early academic partners such as CIRAD, CNRS, and AgroParisTech, and it later expanded through partnerships with universities including University of Oxford, University of Montpellier, and University of São Paulo. Launch milestones involved pilot deployments in Europe, Africa, and Latin America, with dataset contributions from herbaria like the Natural History Museum, London; Royal Botanic Garden Edinburgh; and New York Botanical Garden. Funding and support came from foundations and agencies such as the European Commission, Agropolis Fondation, and national research councils like ANR and NSF, enabling integration with projects at EMBL-EBI, SIB, and the Biodiversity Heritage Library. Technical collaborations involved teams at Google Research, Facebook AI Research, and Microsoft Research on image-recognition benchmarks while publications appeared in journals including Nature, Science, PLOS ONE, and BioScience.

Features and Technology

The platform provides mobile applications for Android and iOS and a web portal, incorporating features such as multi-image observation uploads, metadata capture (date, location, GPS), annotation tools, and taxon pages referencing databases like IPNI, JSTOR Global Plants, and BHL. Core technology uses convolutional neural networks, transfer learning, and ensemble methods developed with partners at ETH Zurich, MIT, Stanford University, and Max Planck Institute for Plant Breeding Research, with training datasets curated from herbaria and field collections at Kew, Smithsonian Institution, and Royal Botanic Gardens, Sydney. The architecture supports APIs and data export compatible with Darwin Core and JSON-LD, enabling integration with platforms such as iNaturalist, eBird, Map of Life, and OpenTree of Life. Ancillary tools include community vetting interfaces, automated quality filters influenced by methods from Google Cloud Vision and TensorFlow, and visualizations using GIS frameworks from Esri and QGIS.

Data Collection and Partnerships

Data collection relies on volunteers, professional botanists, botanical gardens, herbaria, and conservation NGOs including WWF, IUCN, and Conservation International, with specimen-level data cross-referenced to repositories like GBIF, BOLD Systems, and Dryad. Academic partnerships span institutions such as University of California system, University of Helsinki, Universidad Nacional Autónoma de México, and University of Cape Town, while funding and policy collaborations involved UNESCO, European Commission Horizon projects, and national ministries of environment. Regional programs partnered include Kew’s Millennium Seed Bank partnership, the African Plants Initiative, and Brazilian Network of Botanical Collections, facilitating long-term monitoring projects tied to initiatives by Ramsar Convention, CBD, and CITES listings.

Use and Applications

Researchers use the platform for floristic surveys, phenology studies, invasive species monitoring, and ecological niche modeling in conjunction with tools like MaxEnt, R, and ArcGIS, often citing datasets when publishing in journals such as Ecology Letters, Journal of Ecology, and Conservation Biology. Conservationists employ it to support Red List assessments by IUCN and national conservation agencies, while agricultural extension services and forestry departments utilize identifications for pest and disease management in collaboration with FAO and CGIAR centers. Educators and outreach programs deploy the app in schools, botanical gardens, and citizen science curricula developed with institutions like Natural History Museum, Smithsonian Institution, and Royal Horticultural Society.

Reception and Impact

The platform has been recognized in media outlets and scientific reviews alongside projects like iNaturalist, eBird, and Seek for democratizing biodiversity data collection, receiving awards and mentions from bodies including the European Citizen Science Association and national science academies. Its datasets have contributed to peer-reviewed research on global plant distributions, climate change impacts, and biodiversity discovery, cited in analyses by IPBES, Nature Climate Change, and PNAS. Conservation outcomes include improved occurrence records for threatened taxa in regions covered by Conservation International, BirdLife International, and local botanical agencies, influencing policy dialogues at COP biodiversity conferences and regional conservation strategies.

Legal and Ethical Issues

Legal and ethical considerations involve data licensing, intellectual property, and privacy, aligning with frameworks such as Creative Commons, GDPR, Nagoya Protocol on Access and Benefit-sharing, and national biodiversity laws enforced by ministries of environment and agencies like CITES authorities. Debates address data sensitivity for endangered species protection advocated by IUCN and national parks agencies, benefit-sharing with Indigenous peoples and local communities represented by IUCN Members and UN Permanent Forum on Indigenous Issues, and ethical AI use discussed in forums including UNESCO’s Recommendation on the Ethics of AI and academic centers at Oxford Internet Institute and Stanford HAI.

Category:Citizen science Category:Botany Category:Biodiversity informatics