Satellite-based data analysis

Fraunhofer EMI develops technologies and methods for small satellites, addressing payload development, mission implementation, and data products. New approaches for mission concepts in earth observation and science are being developed at the interface between New Space and established space.

Earth observation provides significant contributions to solving global problems in the areas of security, environmental protection, climate change impacts, an enabley new approaches for agriculture, urban development, disaster management, logistics and defense.

Fraunhofer EMI focuses on research and technology development for the acquisition and real-time processing of Earth observation data through satellites.

How do small satellites enable new business models in earth observation? Fraunhofer EMI is developing infrared cameras for small satellites using the New Space approach. Under the motto "validation instead of certification", cost-efficient payloads with commercial components are being developed. These are intended to deliver precise infrared data with a high repetition rate in small satellite constellations. Based on a pilot mission with a thermal infrared imager on the ISS, the EMI spin-off ConstellR is building its own small satellite constellation whose surface temperature data is intended to solve urgent problems such as water management in agriculture.

How can artificial intelligence make on-board data acquisition more efficient? Machine learning enables the analysis of relevant Earth observation targets directly on board. More and more satellites and powerful payloads are generating a growing amount of raw data that is transmitted via ground stations. Fast, automated analysis on board allows relevant targets to be identified and irrelevant data, such as cloud cover, to be sorted out. This reduces the amount of data for the downlink and makes earth observation tasks more efficient.

What difference do commercial FPGA modules make to onboard data processing? Fraunhofer EMI integrates commercial FPGA SoC hardware into system-on-module concepts for highperformance onboard data processing. Modern FPGAs minimize bottlenecks in downlink capacity and enable the analysis of payload data already on board, for example through artificial intelligence. The systems adapt to the specific needs of the payload and meet reliability requirements through redundant design.

How can the resolution of Earth observation images be increased despite physical limitations? Superresolution reconstructs images of higher quality from low-resolution images. To do this, several images are combined or deep learning models are used. The latter use additional training data for feature detection and extraction to determine suitable parameters for resampling. Fraunhofer EMI is investigating super-resolution methods specifically for combining earth observation data from different satellites with different resolutions.

How can new technology and mission concepts be tested quickly in orbit? Thanks to their rapid and cost-efficient implementation, small satellite missions are key to demonstrating new technologies in orbit. They create new application possibilities for which the cost threshold of established systems is too high. Rapid market growth and high innovation potential in the field of small satellite constellations are driving the pace forward. The Fraunhofer EMI is available as a partner to realize new concepts from the sketch to design, implementation and testing to operation.