DEMETER approach towards an agricultural interoperability space

The latest advances of technologies such as IoT, AI and Big Data, among others, have boosted the adoption of smart farming practices that emphasise the use of ICT in farm management processes to exploit the available data. This, however, has led to an explosion of smart farming solutions and data availability that brings new challenges to be addressed. This fact introduces a lack of interoperability between different systems and platforms in the sector, especially the ones offered by different technology providers. Data is usually available by different sources, in different formats, and represented according to different models, thereby hampering data integration and exchange across multiple solutions. The lack of integrated data access and interoperability, in turn, hinders the full potential of value creation based on all data available, and the development of smart services and applications supporting the decision making processes.

The DEMETER project is addressing the need for technical and semantic interoperability through the Agricultural Interoperability Space (AIS) deployed and the underlying Agriculture Information Model (AIM) implemented. The former provides interoperability mechanisms to integrate and deploy a solution, while the latter provides the common (semantic) language used by DEMETER enabled applications to exchange data. In particular, AIM defines the data elements, including concepts, properties and relationships relevant to agri applications, as well as their associated semantics/meaning for information exchange.  Built upon a thorough analysis of the related state of the art and practice, and driven by the elicited stakeholder requirements in DEMETER, AIM aims to establish the basis of a common agricultural data space and enable the interoperability of different systems, potentially from different vendors. This will in turn enable the analysis of data produced by those systems in an integrated manner to make economically and environmentally sound decisions.

AIM design and implementation

AIM has been designed following a layered and modular approach, and is realised as a suite of ontologies implemented in line with best practices, reusing existing standards and well-scoped dominant models as much as possible and establishing alignments between them to enable their interoperability and the integration of existing data. AIM is scalable and can be easily extended in order to address additional needs and incorporate new concepts, maintaining its consistency and compliance. In particular, AIM comprises the following layers:

  • the meta-model layer defining the building blocks of AIM and enabling the back-and-forth conversion between datasets that are based on the property graph model and linked data datasets
  • the cross-domain layer defining relevant concepts and properties that are common across multiple domains, and which enable the interoperability with existing standard models and vocabularies
  • the domain layer defining agri-specific concepts and properties covering different aspects of interest of agri applications, and which enables the integration of relevant vocabularies in the sector.
  • The pilot-specific layer defining additional concepts and properties that are of specific use for particular applications. 
  • Additionally, AIM defines a metadata model that can be used to describe datasets, services or applications in DEMETER.


A key value provided by AIM is that it harmonises and aligns relevant cross-domain standards such as Time Ontology, SOSA/SSN, GeoSparql, QUDT, Data Cubes, with domain-specific models such as Saref4Agri, FIWARE and INSPIRE/FOODIE, bridging various views on the agriculture data and providing a formal representation enabling unambiguous translations between them.

AIM is published as both human and implementation-ready machine-actionable resources, including the formal specifications as ontology modules (OWL ontologies), JSON-LD contexts enabling services to exchange AIM-compliant data based on the already successful JSON format, and SHACL shapes enabling the validation of data against AIM semantics. AIM specification includes guidelines on how to find and identify relevant terms, how to create AIM-based JSON-LD content, as well as instructions to validate the generated content.

In line with FAIR principles [1], the AIM is released using persistent and resolvable identifiers (namely from w3id service), allowing access to the ontology on the Web via its URI, with support for content-negotiation, and ensuring the sustainability of the ontology over time. The AIM URL, by default, redirects and opens the ontology in the OGC definition server.

How will AIM benefit farmers and technology providers?

For farmers, AIM will enable them to use the best suited solution for their needs, including systems and components from different technology providers that will be able to seamlessly interoperate and exchange data. The transparent use of these different components allows farmers to use the best and most cost-effective combination to carry out their activities efficiently and economically, avoiding vendor lock-in. Moreover, having data produced and collected by different systems in an AIM-compliant format will support farmers in their decision making processes, as the underlying tools and analytic services will be able to have an integrated data access to exploit the full value of available data.

For technology providers, on the other hand, producing and consuming data in an AIM-compliant format will allow their systems and components to interoperate with other existing solutions. This will allow them to focus their efforts on developing specialised components reflecting their main expertise, and/or reduce costs, time and efforts needed to develop components that are already available. Also, the possibility to interoperate with components from different providers will allow some providers, especially smaller (e.g., SMEs, start-ups), to enter in otherwise monopolised farming solutions. Additionally, technology providers will be able to ensure the future interoperation with other components, as long as they will be able also to produce/consume AIM-compliant data. 

What will the future of AIM look like?

DEMETER has made a step forward to address the semantic interoperability challenges in the agriculture domain, through the creation of the Agriculture Information Model (AIM) and the proposed approaches to generate AIM-compliant data. As part of the efforts to support its future sustainability and adoption, AIM is starting the process of becoming an OGC recommendation supported by the OGC agriculture Domain Working Group (DWG).  Additionally, several new projects are adopting AIM, to continue its adoption, validation and extension. In fact, not only projects in the agriculture domain can benefit from AIM. For example the project ILIAD which is building a digital twin of the ocean, is reusing the lower layers of AIM and adapting them to the Ocean domain, while the DIVINE project is exploiting AIM to promote the agri data economy.  In many of these projects the goal is not only to produce standard data (i.e., AIM), but also to expose standard APIs to facilitate the data exchange between different systems/tools, such as the RESTFul OGC APIs to provide geospatial data to the web.


[1]  Wilkinson, M.D., Dumontier, M., Aalbersberg, I.J., Appleton, G., Axton, M., Baak, A., Blomberg, N., Boiten, J.W., da Silva Santos, L.B., Bourne, P.E. and Bouwman, J., 2016. The FAIR Guiding Principles for scientific data management and stewardship. Scientific data, 3(1), pp.1-9.