Astronomical observations with smart telescopes produce visually striking but difficult-to-interpret raw sky images without guidance. We present an AI-based object detection approach that identifies satellites, Deep Sky Objects (nebulae, galaxies, star clusters, etc.), and sunspots directly in these images and generates annotated visualizations. In guided educational stargazing sessions with young participants (aged 10–16), raw and AI-processed views are compared in real time, highlighting the importance of expert guidance to interpret what is seen and avoid misidentification. Beyond automation, AI processing acts as a pedagogical trigger, helping beginner observers formulate questions they would not have asked and fostering discussion around observation and data interpretation.

Biography:

As Lead R&T Scientist at Luxembourg Institute of Science and Technology (LIST), Olivier has been supervising space-related research projects for several years (MILAN, MILAN2, NEOD, NEOD2, EASE), funded by LIST, the Luxembourg National Research Fund (FNR) and private companies, while leading work packages of large Luxembourgish and European AI projects. Author of numerous academic papers, Olivier is also member of Program committees of international conferences, and regular reviewer of reputed journals.