MAIN OBJECTIVES & OUTCOMES

AgriHUB is a research group within the Beyond Center of Earth Observation Research of the National Observatory of Athens. We conduct research in the domain of Big Earth data and Artificial Intelligence trying to understand agro-ecosystems and develop applications that support sustainable and resilient agriculture. Over the last five years, our team has participated in more than ten EU-funded projects (see here) providing agricultural information services that utilize satellite data, drone images, street-level images, geotagged photos from the field and numerical weather predictions to address real-world problems in the thematic areas described below.

Through this work, AgriHUB (see team members here) has developed the following novel applications: agrowth (phenology and yield predictions), resagri (weather peril earth warning system), optimal sowing recommendation system (with Corteva) and DataCAP (policy monitoring system). These applications are product of systematic research that has culminated to several scientific publications (see here): in crop classification using satellite data and machine learning, semi-supervised and unsupervised phenology predictions for real-world scenarios, explainable predictions of the onset of pest harmfulness in cotton, fusion of satellite and street-level images for enhanced agricultural monitoring, agriculture monitoring data cubes and distributed AI pipelines in HPC to handle the massive amounts of satellite data, land suitability for applying specific management practices using causal machine learning and assessment of the impact of digital agriculture tools using causal inference

Thematic Areas

• Monitoring of agricultural policies (e.g., Common Agricultural Policy of the EU)
• Food Security
• Ecosystem Services and Climate Change
• Smart and resilient farming

Research Areas

• Climate Change and (Agro)ecosystem Services
  • Understanding the drivers of Earth system change
  • Future trajectories of ecosystem services


• Agriculture Modeling
  • Blending networks and process-based models
  • Blending Earth observations and meteorological data


• Information Extraction from Remote Sensing Images
  • Computer vision on images from heterogeneous sources
  • Big Earth data technologies and distributed learning