© ESA

07.01.2022

Using artificial intelligence and open-source satellite data to ensure food security

Can artificial intelligence and open remote sensing data be used to predict crop yields? Programmers from countries in Africa learn about this in training courses.

By 2050, the world’s population will be around 9.7 billion. Demographic growth and ongoing climate change are making it more and more challenging to produce our food.

New approaches are needed to ensure food security for the global population in future. Open-source and freely accessible satellite images taken from the sky at regular intervals play a crucial role. They provide information that can be analysed automatically and accurately in a matter of seconds using artificial intelligence (AI). These applications can be used as monitoring tools and early warning systems to analyse growing conditions, predict yields and prevent crop failures.

Using and programming the AI applications correctly requires experts with specialist knowledge. On behalf of the German Federal Ministry for Economic Cooperation and Development (BMZ), the Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) GmbH therefore is implementing trainings for AI specialists. In cooperation with the public benefit organisation Radiant Earth Foundation and Makerere University in Uganda their measures range from data competitions and creating open-source training data sets to bootcamps.

Bootcamps: know-how, practical exchange and training data for all

At the first virtual bootcamp for programmers, 40 participants from Ghana, Rwanda, South Africa and Uganda found out how to use machine learning – a branch of AI – to analyse satellite images automatically. This includes calculating how much wheat or maize is planted in an entire province, for example.

Lillian Muyama was one of the participants: ‘The course gave me an in-depth insight into this area. It exceeded my expectations,’ she says. Lillian, a qualified data scientist from Uganda, was excited about the use of AI in sustainable agriculture right from the start:

‘The use of satellite data has huge potential, especially in Africa. And not just for agriculture but also for civil protection, for example.’ Measures like the international bootcamp are important, she adds, because the necessary technical expertise and access to regional training data sets have been lacking up to now. Muyama sees improved access to data and exchange with colleagues as important steps to drive forward the development of the AI sector in Africa.

Lillian wants to use her newly acquired expertise in her job and put it into practice as soon as possible – for instance for comparing sensor readings with air quality data from satellite images: ‘We’re not quite there yet, but satellite data will certainly be part of my work in future. We are already planning to use the initial data sets for projects in 2022.’  

All training participants are encouraged to pass on their knowledge. For this, the lectures and training material has been made openly available for everyone on the e-learning platform atingi.

Further Information