STELAR Project

Stelar logo
Menu

Pilots

Pilot A: Risk prevention in food supply lines

  • Pilot Leader: AGROKNOW
  • Participants: ATHENA, TU/e, UniBwM, FSH
  • Technology used: LLM-based Named Entity Resolution, pyJedAI-Entity Matching, pyJedAI-Schema Matching, Correlation Detective

PILOT A Checklist

  • Risk prediction from incident data
  • Time-series updates for evolving threats
  • NLP for hazard identification
  • Ingredient mapping with LLMs
  • Data harmonisation and schema alignment
Basket with vegetables
The food system is rapidly changing, becoming increasingly digitized. Diverse data is being generated by all entities (farmers, food companies, national authorities, consumers), including farm management data, weather data, environmental data, recalls data, border rejections data, laboratory testing data, regulation data, production data, trade data, price data, social data, product reviews data, inspection data, country risk data, audit data and others. These are stored in local databases in different formats and languages, resulting in disconnected data silos. To perform a complete and timely risk assessment, understand emerging risks, and protect the consumers, domain experts need to be able to find, combine, process and extract meaning from as much data as possible across the entire supply chain. This use case will focus on how advanced data management technologies can be used to enable the generation of integrated, AI-ready datasets for food risk prediction.

Pilot B: Early crop growth predictions

  • Pilot Leader: Vista
  • Participants: ATHENA, TU/e, UniBwM, FSH
  • Technology used: satellite imagery, EO data fusion, field segmentation, time series imputation, crop classification

PILOT B Checklist

  • Crop monitoring
  • Early yield prediction
  • Optimisation of agricultural practices
  • Risk management
  • Supporting sustainability
Plant growing from the ground

Predicting crop growth and yield development is crucial both on the local level for farm management measures, as well as on the regional to continental level to predict food supply and allow early warnings for food shortages due to yield loss. Earth Observation (EO) data combined with modeling are a crucial input for these services, as they allow for scalable and continuous service delivery from an independent source. However, for farmers, existing predictions come too late in the season. To effectively influence crop development with smart farming measures, predictions need to happen while the crops are still in the vegetative phase, i.e., much sooner in the season. In this use case, we will combine data from different sources (e.g., operational multispectral and SAR satellite data as well as available hyperspectral data) to provide the needed input for physical modeling and deep learning techniques that derive crop type as well as crop status and current crop growth, toward the goal of predicting crop development as early as at vegetation start in spring.

Pilot C: Timely precision farming interventions

  • Pilot Leader: Abaco Group
  • Participants: ARC, UoA, FSH
  • Technology used: PILOT C Technologies

  • Data management tools:

    • Dataset search
    • Entity matching
    • Geospatial analytics
    • Time series imputation
    • Predictive analytics

    Types of data:

    • Weather Station Networks
    • Digital Elevations Model
    • Soil Maps

PILOT C Checklist

  • Timely precision farming interventions
  • Integration of EO and on-farm data
  • Decision support for farm planning
  • Localised and adaptable digital tools
  • Promoting sustainability and traceability

The integration of EO data with farming technologies selected and contextualized according to local specificities can empower end-users with valuable data to support land management and crop planning. The management of water, fertilizers and pesticides must satisfy requirements that software helps to manage. This aids users in developing precision farming agricultural strategies. Prescription maps are produced for variable rate treatments or precision irrigation. Correlating best practice activities, meteorological indicators, and IoT data from crop proximal sensing and food processing equipment, can offer information on how a crop product may have been contaminated or “simply” dirtied by soil splashing. Then, the food processing equipment can use this data to manage water and other resources needed accordingly. In this use case, ​​we will integrate and correlate data from the crop, with the post-harvest stand-alone machinery to enable the creation of a data-driven decision support system for addressing many different issues that influence the plant status in field and in post-harvest.

Man holding a tablet in his hand

STELAR Project
Powered by  GDPR Cookie Compliance
Privacy Overview

This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.