Darker Skies, Deeper Insights: Event Cameras at the Bavarian Forest International Dark Sky Reserve

Event cameras are an emerging sensing modality for space applications due to their ability to detect small changes in brightness. Rather than capturing full image frames, event cameras generate events when the intensity of light at an individual pixel increases or decreases. Despite their promise, event cameras remain a relatively new technology, meaning that the number of astronomy-specific tools and available recordings of the night sky is very limited.

To address this gap, Dr. Sydney Dolan, Jaspar Sinderman, Lara Schuberth, Rugved Arge, and Gordon Hahmeyer recently completed an observation campaign at the Bavarian Forest International Dark Sky Reserve. Welcome by Professor Dr. Detlef Koschny, the team worked together to set up two event cameras attached to telescopes to record the night sky. The Bavarian Forest Dark Sky Reserve offers a one-of-a-kind astronomy opportunity, as it is a protected region designated to minimize artificial light pollution. Light pollution is caused by excessive or poorly directed artificial lighting from sources such as streetlights or buildings. This undirected light scatters in the atmosphere and creates a diffuse sky glow that washes out faint star signals. Compared to urban environments like Munich, which have a diffusive sky glow, the dark sky reserve has significantly darker and clearer skies. The quality of these skies allows for exciting opportunities to observe stars, satellites, and other dim objects that would remain undetectable. 

Over three nights of observation, the campaign yielded a diverse dataset of satellite passes, rocket body detections, and star recordings. Multiple telescope configurations were tested, including variations in mount tracking speeds and optic lens. Diversity in speed of the ego camera when using an event camera is important. Movement of the camera can help stagnant objects, like stars at infinity, generate enough contrast to send enough light to the pixel and generate an event. 

Initial post-processing of event recordings revealed the event camera can detect objects with apparent visual magnitudes as faint as 11. This sensitivity exceeds that of the human eye by several magnitudes and is beyond the theoretical projections produced by our group. Notably, this level of detection was achieved within fraction of a second. Unlike conventional frame-based cameras, which require longer exposure times to accumulate enough light for faint objects to appear in an image, event cameras response immediately to small changes in brightness. This allows them to capture faint signals much more quickly.

A notional astronomy processing pipeline has been developed to label and catalog stars within the dataset, providing additional insight into system performance. Beyond these initial results, the dataset represents a valuable resource for future research at the Chair of Spacecraft Systems, where it will support the development and validation of advanced event-based algorithms for satellite detection and tracking.

Relevant Publications

• Sydney Dolan, Lara Schuberth, Rugved Arge, Ramón María García Alarcia, Vincenzo Messina, Clemente J. Juan Oliver, Francesco Salmaso, Jaspar Sindermann, Federico Sofio Avogadro, Alessandro Golkar, “Design and Analysis of an Event Camera Payload for Space-Based Object Detection on the EventSat 6U CubeSat Mission.” 15^th IAA Symposium for Small Satellites for Earth System Observation, 2025.
• Sydney Dolan, Alessandro Golkar. “Sun-E: Dataset and Benchmark for Event-Based Sun Sensing.” IEEE Winter Applications of Computer Vision, 2026.
• Lara Schuberth, Sydney Dolan, Alessandro Golkar. “Neuromorphic Vision Sensor Noise Characterization and Detection for CubeSat Technology Demonstration Mission.” 76^th International Astronautical Congress, 2025.