As part of an international cooperation between the DHBW Ravensburg Campus Friedrichshafen and the Abdelmalek Essaadi University Faculty of Sciences and Techniques of Tangier, SeeSat e.V. is working with students on the development of an AI-supported system for in-orbit image labeling as part of the ERWIN mission. The aim of the project is to identify and prioritize satellite images of potential sources of fire at an early stage so that they can be transmitted to the ground station with greater urgency.

Over the course of the project, the students from Tangier will develop model concepts, analyze suitable AI frameworks and implement prototypes for training environments. They are also creating and collecting training data sets in order to train the AI models specifically to detect fires. This autonomous image processing will significantly increase the efficiency of the ERWIN mission, as critical environmental information will be made available more quickly.

In the long term, the project aims to produce a fully integrated, precise and autonomous system for in-orbit image processing. This will enable reliable detection of fires and prioritized data transmission to ground stations, optimizing response time to environmental disasters and improving environmental and disaster protection.

We are also working together with the University of Hagen on autonomy concepts for the ERWIN mission. In contrast to machine learning for in-orbit image labeling, expert systems are used here. These allow the use of expert knowledge and telemetry data to control and troubleshoot the satellite.