Join the 2025 Geo-AI Challenge: Mapping Cropland in Arid Regions with AI

The 2025 edition of the Geo-AI Challenge is now live! Organized in collaboration with the International Telecommunication Union (ITU) through its AI for Good platform – and hosted on the Zindi platform, which connects a global community of data scientists – the competition calls on data scientists, researchers, and AI practitioners to contribute innovative solutions for cropland mapping in dryland regions using artificial intelligence (AI) and machine learning (ML). 

Accurate cropland maps are essential for monitoring food production, managing natural resources, and planning agricultural interventions. However, cropland mapping in arid and semi-arid regions remains a significant technical challenge due to the difficulty of distinguishing cropland from pasture and steppe using traditional remote sensing methods. 

This year’s challenge focuses on two complex landscapes: Fergana (Uzbekistan) and Orenburg (Russia), where distinguishing cropland from pasture and steppe presents a unique technical hurdle. Participants are tasked with developing cost-effective and accurate classification methods using time-series satellite imagery and AI/ML techniques. 

By participating, researchers and practitioners will contribute to the advancement of global cropland mapping, promoting a more precise and comprehensive understanding of agricultural landscapes. 

Read more about the evaluation process of the challenge, the prizes, the timeline and how to participate here. 

Competition closes on 29 September 2025.