Scientists from IT4Innovations are working on the European Space Agency's new AIOPEN project using artificial intelligence

The world's population continues to grow, and an increasing number of people live in urban areas. In 1960, only about one-third of the world's population, i.e., 1 billion people, lived in cities. Today, it is more than half of the world's population, approximately 4.5 billion people. And this trend is increasing. According to United Nations projections, the urban population will have almost doubled by the end of 2100, approaching 9 billion.

IT4Innovations has been working on the topic of global urbanisation, supported by a collaboration with the European Space Agency (ESA), since 2015. The first ESA project in which IT4Innovations was involved was Urban TEP, which uses data storage infrastructure and provides data processing services. The project, led by the German Aerospace Center (Deutsches Zentrum für Luft- und Raumfahrt, DLR), has created a thematically oriented platform to facilitate the search for Earth observation data. The platform offers, for example, the generation of cloud-free images of the Earth's surface, allowing an unobstructed view of built-up areas and tracking the evolution of global urbanisation.

In 2019, IT4Innovations began working with Space Applications Services and others on the BLENDED project. Together, they prototyped a distributed peer-to-peer platform using machine learning to analyse space data. IT4Innovations was responsible for creating a deep learning model that processed remotely sensed Earth observation data to analyse and predict the growth of urban agglomeration in the monitored cities.

In 2023, ESA also supported another project - AIOPEN, involving IT4Innovations in collaboration with Space Applications Services, Telespazio, KP Labs, and SERCO. The project aims to create an AIOPEN platform that allows users, regardless of their technical expertise, to use and apply artificial intelligence to process remote sensing and satellite imagery data. It will provide visual workflows and pre-built templates that users can customize with available Earth observation data. The platform will have a user-friendly interface, and the tools will be intuitive.

The AIOPEN platform will be created by combining and extending the ASB, EOPEN, and EOEPCA platforms with new, innovative services based on artificial intelligence and machine learning. It will be hosted in SERCO's ONDA DIAS and fully capable of processing tasks in remote environments.

Two case studies will be implemented on the platform during the project: Urban Change Detection with Transformer Architecture, for which IT4Innovations is responsible, and Deforestation Tracking.

Cities are dynamically changing places that are at the centre of modern life. Yet, they also face climate change, pollution, and urban sprawl challenges. The Urban Change Detection case study aims to use machine learning models to monitor urban growth, thereby enabling comprehensive regional development planning and preventing unwanted urban sprawl. Earth observation data and deep neural networks (DNNs) will be used to detect urban change, and a digital twin of Earth's (urban) changes will also be created. Artificial intelligence will allow satellite imagery to be analysed in depth. It will play a key role in detecting changes in urban areas over time, whether to identify urban expansion zones or to support urban planning with an emphasis on sustainable growth.

Another long-standing problem in many countries is deforestation, which the AIOPEN team will also address under the leadership of KP Labs. Monitoring, segmentation, and analysis of forests will also be carried out using artificial intelligence, algorithms, and publicly available data.

Václav Svatoň, project investigator for IT4Innovations, adds: “The AIOPEN project aims to extend existing commercial platforms with AI functionalities. The project's results, i.e., the provision of services enriched with innovative AI functionalities, will be published in the Network of Resources (NoR).”