Cos4Cloud: Strengthening Citizen Science Through Open Data and Artificial Intelligence
Conducted from November 2019 to February 2023, the European project Cos4Cloud developed new digital services to improve the collection, quality, and sharing of citizen science data. Pl@ntNet played a central role by making its plant identification technologies available to other scientific and citizen communities.
Every day, thousands of people observe nature, photograph species, or monitor the quality of their environment. Thanks to citizen science platforms, these contributions form valuable datasets for research, biodiversity conservation, and environmental change monitoring. However, their use at scale raises several challenges: how to verify the quality of observations, link different platforms together, facilitate data reuse, and properly acknowledge the work of contributors?
Funded by the European Union’s Horizon 2020 framework program, Cos4Cloud (“Co-designed Citizen Observatories Services for the European Open Science Cloud”) brought together fifteen European and international partners to address these challenges. The project designed and co-tested thirteen digital services with end users, relying on artificial intelligence, automated image recognition, data analytics, and open standards. These tools cover biodiversity observation as well as air and water quality monitoring.
Pl@ntNet at the Core of the Project
Within Cos4Cloud, the Pl@ntNet team led the development of services dedicated to data collection, processing, and quality control. One of the key outcomes is the Pl@ntNet-API service, which allows scientists, developers, and other citizen observatories to integrate Pl@ntNet’s plant identification engine directly into their own applications.
Using one or several photographs (leaf, flower, fruit, or the entire plant), this service returns the most likely matching species. It can be tailored to a specific flora or geographic region, thereby facilitating the creation of specialized applications—for example, to monitor plant life across a specific area, support a research program, or rapidly detect invasive species.
The project also helped to:
- Facilitate research access to Pl@ntNet observations;
- Improve interoperability across different citizen science platforms;
- Build curated, verified image datasets to train and benchmark AI models;
- Develop methods combining computer vision with geographic and environmental data;
- Promote data sharing aligned with the FAIR principles: making data Findable, Accessible, Interoperable, and Reusable.
A Virtuous Circle Between Citizens, Scientists, and Artificial Intelligence
Cos4Cloud is built around a simple idea: citizen-generated observations help scientists and train automated identification algorithms. In turn, artificial intelligence and expert scientific curation make it easier to identify and validate incoming observations.
This virtuous circle lies at the very heart of Pl@ntNet. The larger, more diverse, and accurately identified the community contributions are, the more powerful the recognition models become. These improved models then guide an ever-growing community of users in discovering the plant life around them, generating new, high-value data for biodiversity knowledge.
By bridging citizen science, artificial intelligence, and open science, Cos4Cloud helped establish Pl@ntNet not only as a high-performing plant identification app, but also as a shared research infrastructure and a reusable service for external initiatives.
- Duration: November 2019 – February 2023
- Program: Horizon 2020
To learn more:
