Tools and methods for extended plant PHENotyping and EnviroTyping services of European Research Infrastructures

Project Information

PHENET

Grant agreement ID: 101094587

DOI

10.3030/101094587

EC signature date 1 December 2022

Start date 1 January 2023

End date 31 December 2027

Funded under

• Research infrastructures

Total cost

€ 10 237 131,26

EU contribution

€ 9 993 469,00

Coordinated by

INSTITUT NATIONAL DE RECHERCHE POUR L'AGRICULTURE, L'ALIMENTATION ET L'ENVIRONNEMENT
France

Periodic Reporting for period 2 - PHENET (Tools and methods for extended plant PHENotyping and EnviroTyping services of European Research Infrastructures)

Reporting period: 2024-01-01 to 2025-04-30

Summary of the context and overall objectives of the project

In PHENET, the European Research Infrastructures on plant phenotyping (EMPHASIS), ecosystems experimentation (AnaEE), long-term observation of interactions between nature and people (eLTER) and data science (ELIXIR) join forces to co-develop, with innovative companies, new tools and methods to help identify future-proofed agroecosystems to face their burning challenges.
While RIs are already well equipped with advanced, instrumented sites, PHENET is developing new services to expand access to enlarged sources of in-situ (on farm, in natura…) phenotypic and environmental data. These support the evaluation of agro and ecosystems by leveraging big data strategies relying on (i) new IoT multi-sensor devices and algorithms that capture a variety of phenotypic and environmental data (ii) unleashed access to high-resolution Earth Observation data and (iv) next-generation modelling solutions powered by AI and digital twins.
These developments are tested and implemented across 8 Use Cases (UC) covering a large range of agro and ecosystems in order to demonstrate the wide applicability of solutions. A large effort is devoted to training RI staff and beyond through a collection of training material. Outreach activities aim at enlarging the range of RI users and engaging stakeholders. PHENET also aims to impact the development of innovative companies on phenotyping, envirotyping and precision agriculture as well as the emergence of « future proofed » crop varieties and innovative practices fitted to climate change and agroecological transition.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

PHENET works with industry to develop low-cost, AI-powered & eco-friendly devices equipped with sensors to monitor diverse traits, environmental conditions & key agroecosystem processes—especially those linked to agroecology and climate change. In RP2, WP2 built several systems & developed algorithms for phenotyping soil health, crop diseases, and orchard conditions. These tools were tested in real-world farm and field experiments. The resulting large-scale dataset will now be used to build models that extract valuable phenotypic insights. PHENET will boost the use of satellite imagery for phenotyping and envirotyping. We have developed an automated method to map intra-field structural variability using multi-year data from Copernicus Sentinel satellites. This tool will support RIs by enhancing experimental design and improving statistical analysis in field trials. Beyond research, the method could have broader applications at farm level, enabling the assessment of crop performance variability and the impact of innovative practices over time. Work is now progressing on harmonization and data fusion techniques to fully exploit EO data. A comprehensive EO image archive has been compiled to support this effort.
The PHENET Open Science services has capitalized on existing data standards, databases, storage & compute services. They have been deployed, connected together and for some of them improved in the frame of the project.
To fully exploit the data, we concentrated on the design & development of hybrid models and 3D-digital twins, and the development of services for selected UCs. Results are promising, and developed methods have been published in peer reviewed articles & open source libraries.
PHENET has been focusing on upskilling the community both on technological aspects and knowledge around increasing the attractiveness, quality and impact of training activities. Business-related stakeholders were addressed. UCs are the core of PHENET activities and all have been successfully implemented and benchmarked tools and methods co-developed with the technical WPs on their specific scientific questions. These tools and methods represent potential services that may be part of the service portfolio of the RIs involved in PHENET.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

PHENET has delivered several breakthroughs that go beyond the current state of the art in phenotyping and envirometrics. Significant progress has been achieved in the development and deployment of AI-powered embedded devices, including soil probes, crop rovers, connected sticks, and orchard phenomobiles. These systems now enable real-time monitoring of a broad range of agroecological traits, with the first successful implementations of embedded AI inference chains for in-field data pre-processing. In parallel, novel EO tools have been developed, notably an automated method for mapping within-field structural variability over multi-year satellite archives—a first at the European scale—enhancing trial design and statistical robustness. The creation of digital twins combining 3D+t plant architecture simulation and hybrid AI/process-based models paves the way for advanced prediction of genotype × environment × management interactions. These tools have been made available as open-source libraries. Data interoperability has been significantly advanced through harmonized standards (MIAPPE, ICASA, Glosis) and the deployment of multi-use PHIS database instances. Several deep learning models now outperform human visual assessment. UCs have demonstrated the adaptability of PHENET tools across diverse agroecosystems and farming practices, supporting their transformation into transferable services for Research Infrastructures. These results mark a clear step-change, enabling distributed, low-cost, automated, and interoperable phenotyping and envirotyping at landscape scale.

image-for-rp2.png