Deep learning solution for the automatic detection of objects from space

In the context of the growing number of satellites that are placed weekly into low Earth orbits (LEO) around the Earth, astronomical images are becoming increasingly contaminated by these objects and other orbital debris, crossing the field of view. Since manually sorting the frames that can be processed for scientific purposes is not a viable solution, a group of researchers from the Astronomical Institute of the Romanian Academy proposes an alternative based on deep learning techniques. The custom model, trained on a consistent set of all-sky images acquired at the Berthelot Observatory (AIRA), achieves 100% precision and 91% sensitivity, meaning that only 9% of objects (satellites or orbital debris) could not be detected, while all detected objects were correctly identified.

From an operational perspective, this combination of high sensitivity and the absence of false alarms is highly advantageous: frames flagged as “contaminated” are sent directly to the SST (Space Surveillance and Tracking) analysis pipeline, while “clean” frames enter the scientific data-processing pipelines without additional manual inspection.

The article published in Astronomy and Computing can be accessed here: https://doi.org/10.1016/j.ascom.2026.101081

Visualization of detection predictions on image data
Exo4Edu teacher training activities in Bucharest and Cluj-Napoca, 14–15 February 2026

Exo4Edu teacher training courses, 14–15 February 2026

Over the weekend of 14–15 February 2026, a new session of teacher training courses took place as part of the Exo4Edu programme, delivered through the collaboration of the teams from Bucharest and Cluj-Napoca. The course activities were designed as a coherent learning pathway, combining scientific explanations with hands-on exercises and teamwork. In this way, the participating teachers received information and developed the skills needed to turn their students’ curiosity about exoplanets into concrete classroom activities, based on up-to-date educational resources.

The programme began with a lecture on extrasolar planets in their historical and astronomical context, followed by discussions and questions. The teachers then moved on to the practical activity “How do you imagine an exoplanet?”, an exercise that stimulated both imagination and reasoning by drawing on accessible scientific information. The courses continued with a presentation dedicated to exoplanets and detection methods, followed by a hands-on activity using the Stellarium software. Continuing in the same practical mode, participants were able to observe, through specialized equipment, the star closest to us – the Sun – during which the general operating principle of the telescope was presented.

The second day of the course began with a visit to the Astronomical Institute of the Romanian Academy. This moment provided a special setting in which science naturally met historical memory. The group of trainees followed in the footsteps of the great forerunners who, at the Bosianu Villa, helped prepare the Act of the Union of the Romanian Principalities of 24 January 1859. During the guided tour, the teachers were able to see in the Institute’s museum rooms, the scale of the national effort to build and capitalize on a vast scientific heritage, which made it possible to achieve and sustain standards of excellence in Romanian astronomy.

As the course continued, attention shifted to the project’s resources and to teaching scenarios that are easy to replicate in the classroom. Participants worked in groups with the Exo4Edu platform, explored the Exo3D game and the exoplanet database, which enabled them to extract lists of exoplanets and use them in educational contexts. The applied component continued with elective activities, such as developing an exoplanet exploration scenario grounded in information and plausible reasoning.

The session concluded with a group-wide reflection on the topics covered, aimed at identifying best practices and strategies that can be readily adapted in schools, as well as with the final evaluation and the awarding of participation certificates. The courses were delivered by educators, lecturers, and trainers Mirel Birlan, Ana Naghi, Iharka Csillik, Carmen Busu, Alin Nedelcu, Cristian Omăt, and Sorin Marin.

APSIS: Automated Photometric Survey of Inactive Satellites

Researchers from the Astronomical Institute have recently published a new study presenting APSIS, an automated program for the long-term, high-precision photometric monitoring of inactive satellites in geosynchronous (GEO) and medium Earth orbits (MEO). Using the infrastructure at Berthelot Observatory, APSIS delivers a unique multi-year dataset of satellite lightcurves and astrometry, enabling systematic investigations of rotational dynamics and long-term orbital stability, key components of Space Situational Awareness (SSA).

The study demonstrates that APSIS can detect subtle, near-annual variations in satellite rotation periods, attributed to solar radiation pressure (SRP) torques, and constrain satellite spin-axis orientations through physics-based modeling. In addition, the authors introduce a novel lightcurve inversion framework that reconstructs satellite geometry from unresolved optical observations, providing new insight into the physical properties of derelict space assets. This work highlights the growing role of automated optical surveys in monitoring inactive satellites and improving the safety and sustainability of the near-Earth space environment.

The results have been published in the journal Acta Astronautica.

Real lightcurves and synthetic model

Satellite shape reconstruction
Conducting the activity at the Astronomical Observatory in Cluj. Credit: © Observatorul Astronomic Cluj-Napoca

New activities within the European project Exo4Edu


The Astronomical Observatory in Cluj hosted an educational event on Wednesday, 10 December 2025, organized within the European project “Exoplanets for Education” (2024-1-FR01-KA220-SCH-000256856) Exo4edu, funded by the European Commission through the Erasmus+ Programme. The activity was carried out by the Astronomical Observatory of the Romanian Academy, Cluj-Napoca Branch, the Astronomical Institute of the Romanian Academy, and Babeș-Bolyai University.


The event, initiated and coordinated by Lecturer PhD Iharka Csillik, researcher at the Astronomical Observatory of the Romanian Academy, Cluj-Napoca Branch, brought together students from the Faculty of Mathematics and Computer Science of Babeș-Bolyai University and seventh-grade pupils from a school in Sălaj County. The participants attended a presentation of the Exo4EDU application, followed by a practical simulation conducted by the university students.


The initiative made use of experiential learning methods focused on the study of exoplanets. In the first part, topics from the field of astronomy were presented in an accessible and applied manner, aiming to clarify new concepts for the pupils. Subsequently, an application dedicated to interstellar travel was used, on the basis of which 12 mixed groups were formed. Each participant selected an exoplanet from the exoplanet.eu database and, with the support of the students, estimated the time required to reach it, depending on the speed of the fastest spacecraft. The activity concluded with astronomical observations using the telescope, targeting the planet Saturn, the star Vega, and the open cluster Pleiades, thereby contributing to stimulating pupils’ interest in astronomical research.

Credit: © Observatorul Astronomic Cluj-Napoca

News

Latest news from A.I.R.A.

Read more

Astronomical data

Ephemerides

Astrodata

Research

Our latest publications

Research

Careers

Join us for a career in astronomy.

Join us