AI4EO (artificial intelligence for Earth observations) is a European Space Agency (ESA) initiative from the Φ-lab of ESA’s Directorate of Earth Observation Programmes, striving to bring the worlds of AI and EO closer together by encouraging interaction and collaboration between the two communities. Each year, AI4EO hosts various AI-centric “challenges” wherein participants are challenged to develop new machine learning (ML) and AI techniques to solve a particular task using Earth observation data. This year’s “#AI4FoodSecurity” challenge was the program’s first multi-organizational effort, with Planet, TUM, DLR, and Radiant Earth teaming up with ESA to encourage AI solutions that support global food security. The challenge focused on satellite data applications for agricultural monitoring, with participants setting out to innovatively apply Planet Fusion and Sentinel imagery for crop identification. At the end of each challenge, AI4EO hosts an award ceremony to announce the winners and present their solutions, with this year’s keynote presentation given by NASA Harvest AI Lead, Dr. Hannah Kerner.
Dr. Kerner has extensive experience in developing machine learning solutions for both planetary and Earth applications, with her current work at NASA Harvest - and in close partnership with Planet - focused on in-season crop type classification systems around the world. She congratulated each of the challenge finalists on their creative solutions, including winners TCSA-AI, and spoke of the impacts that this work has across the agricultural sector. Dr. Kerner emphasizes that “…our work does not stop here. The end goal is not the test metric, but rather the realization of the benefit of these technologies for people, and our non-human co-inhabitants, across our fragile planet. We need to invest in creating AI solutions that work for real-world problems, real-world data, and real-world people.”
Applying AI methodologies for improving crop monitoring techniques enables us to better understand where crops are growing, how well they are growing, and how much is growing. The ability to do this worldwide on a regular basis is critical to knowing how much food is available for global consumption and supports transparency and market stability. The compute power afforded by AI and ML far outperform the otherwise manual effort required to classify cropped areas around the world, which is where this technology is incredibly valuable.
AI4EO emphasizes that “Through these challenges we will foster the growth of the AI4EO community, support researchers and coders by promoting their work and use AI to extract more information from EO to solve some of the pressing challenges faced by our society.”
Visit AI4EO, ESA Φ-Lab, and Planet to learn more about these challenges and how to get involved.