EO4FoodSecurity: Using Earth Observation Enabled Land Cover Classification for Characterizing Global Food Security on Regional Scales

Introduction

Characterizing the state of global food security is essential in devising and evaluating policies and programs for effective decision making. The concept of food security is multidimensional and dynamic and is often compounded by the challenge of obtaining relevant data.

Challenges in Food Security Measurement

Moreover, finding appropriate indicators that specifically encompass the four dimensions of food security (including physical availability of food, economic and physical access to food, food utilization, and sustainability) as specified by UN FAO remains a challenging task. There exists a variety of different measures for assessing the food security situation, but they merely focus on nutrition and physical aspects and thus provide incomplete assessments related to the problem.

Project Objectives

In this PoC project, I aim to extend the unique AI algorithms and the big EO data management features developed in the ERC StG “So2Sat” to characterize the state of global food security on regional scales using multimodal data derived from satellite imagery and auxiliary open data, and offer our software as a commercial, integrated service.

Business Case Development

Within the PoC, a comprehensive business case that will assist us in designing an exploitation strategy will be developed. Achieving these objectives will augment the capability of our existing AI solution for land cover/land use mapping to infer the crucial aspects of food security and sustainability.

Value Proposition

Our value proposition in EO4FoodSecurity is a set of professional solutions to extract relevant indicators for characterizing food security by retrieving them from big EO data and other open sources using AI.

Example Applications

1. Generating land use maps.
2. Using land use maps along with other information extraction modules of So2Sat (such as population density, road, and building footprints).
3. Integrating other open data (e.g., meteorological, nutrition) to generate food security maps at unprecedented finer spatial and temporal scales.

We aim to support these solutions in an easy-to-use, interactive big EO data analysis platform.