Ph.D. in Astronomy, University of St. Andrews, 1995
RESEARCH, TEACHING, or OTHER INTERESTS
Astronomy and Astrophysics, Physics and Astronomy, General Physics and Astronomy
31
Scopus Publications
Scopus Publications
Multispectral YOLO: generic feature fusion framework for solar active region detection António Santos, Filipa S. Barros, J. J. G. Lima, Rui F. Pinto, André Restivo, Luís F. Teixeira Machine Vision and Applications, 2026 Monitoring solar phenomena, such as sunspots and active regions, is crucial for ensuring astronaut safety, telecommunications reliability, and predicting terrestrial events like auroras. Traditional methods for detecting these phenomena have limitations in accuracy and baseline maintenance. This paper presents a novel deep learning object detection method that leverages multispectral image data from satellites to enhance the detection of "sunspots" and active regions. Utilizing images from the SDO satellite and annotations from the DeepSDO dataset, we constructed a new dataset composed of aligned observations from HMI Ic, AIA 211 Å, and AIA 335 Å. We adapted and developed a stock YOLOv5-based model capable of handling and fusing any number of input images. Two fusion methodologies, early and late fusion, and three different fusion modules—CatFuse (simple concatenation), CBAMC (CBAM-based module), and TransEnc (transformer encoder)—were implemented and tested. Statistically analysing the results via the Friedman test ( p =0.05) revealed significant performance differences among the evaluated models, which were confirmed through pairwise Wilcoxon post-hoc tests. From the approaches tested, CatFuse with early fusion achieved significantly higher detection performance than the other models, with a mAP@0.5:0.95 of 0.52 and a mAP@0.5 of 0.94, an improvement of $$\\Delta $$ =0.02–0.36 for mAP@0.5:0.95 and $$\\Delta $$ =0.02–0.28 for mAP@0.5, depending on the baseline model. This result was marginally better than the best baseline (YOLOv5 with a single HMI image) and comparable to other state-of-the-art models, demonstrating a modest but consistent improvement of multispectral image fusion for this task.
Using Recurrent Neural Networks to improve initial conditions for a solar wind forecasting model Filipa S. Barros, Paula A. Graça, J.J.G. Lima, Rui F. Pinto, André Restivo, Murillo Villa Engineering Applications of Artificial Intelligence, 2024 Solar wind forecasting is a core component of Space Weather, a field that has been the target of many novel machine-learning approaches. The continuous monitoring of the Sun has provided an ever-growing ensemble of observations, facilitating the development of forecasting models that predict solar wind properties on Earth and other celestial objects within the solar system. This enables us to prepare for and mitigate the effects of solar wind-related events on Earth and space. The performance of some simulation-based solar wind models depends heavily on the quality of the initial guesses used as initial conditions. This work focuses on improving the accuracy of these initial conditions by employing a Recurrent Neural Network model. The study’s findings confirmed that Recurrent Neural Networks can generate better initial guesses for the simulations, resulting in faster and more stable simulations. In our experiments, when we used predicted initial conditions, simulations ran an average of 1.08 times faster, with a statistically significant improvement and reduced amplitude transients. These results suggest that the improved initial conditions enhance the numerical robustness of the model and enable a more moderate integration time step. Despite the modest improvement in simulation convergence time, the Recurrent Neural Networks model’s reusability without retraining remains valuable. With simulations lasting up to 12 h, an 8% gain equals one hour saved per simulation. Moreover, the generated profiles closely match the simulator’s, making them suitable for applications with less demanding physical accuracy.
Leveraging Physics-Informed Neural Networks as Solar Wind Forecasting Models Nuno Costa, Filipa S. Barros, João J. G. Lima, Rui F. Pinto, André Restivo Esann 2024 Proceedings 32ndeuropean Symposium on Artificial Neural Networks Computational Intelligence and Machine Learning, 2024 Space weather refers to the dynamic conditions in the solar system, particularly the interactions between the solar wind -a stream of charged particles emitted by the Sun -and the Earth's magnetic field and atmosphere.Accurate space weather forecasting is crucial for mitigating potential impacts on satellite operations, communication systems, power grids, and astronaut safety.However, existing solar wind coronal models like MULTI-VP require substantial computational resources.This paper proposes a Physics-Informed Neural Network (PiNN) as a faster yet accurate alternative that respects physical laws.PiNNs blend physics and data-driven techniques for rapid and reliable forecasts.Our studies show that PiNNs can reduce computation times and deliver forecasts comparable to MULTI-VP, offering an expedited and dependable solar wind forecasting approach.
Nonradial and nonpolytropic astrophysical outflows XI. Simulations of the circumstellar environment of RY Tauri C. Sauty, R. M. G. de Albuquerque, V. Cayatte, J. J. G. Lima, J. F. Gameiro Astronomy and Astrophysics, 2022 Context. Recent observational evidence has shown that RY Tau may present two different outflow stages, a quiescent and a more active stage. We try to model this phenomenon. Aims. We have performed new 2.5D magnetohydrodynamical simulations of the possible accretion-outflow environment of RY Tau based on analytical solutions with the aim to reduce the relaxation time. Methods. We used the analytical self-similar solution that we used to model the RY Tau microjet as initial conditions. In the closed field line region of the magnetosphere, we reversed the direction of the flow and increased the accretion rate by increasing the density and velocity. We also implemented the heating rate and adjusted it according to the velocity of the flow. The accretion disk was treated as a boundary condition. Results. The simulations show that the stellar jet and the accreting magnetosphere attain a steady state in only a few stellar rotations. This confirms the robustness and stability of self-similar solutions. Additionally, two types of behavior were observed that are similar to the behavior observed in RY Tau. Either the steady stellar outflow and magnetospheric inflow are separated by a low static force-free region or the interaction between the stellar jet and the magnetospheric accretion creates episodic coronal mass ejections that originate from the disk and bounce back onto the star. Conclusions. The ratio of mass-loss rate to mass-accretion rate that coincides with the change in behavior observed in RY Tau lies within the range of ratios that have been measured during the period in which the initial microjet was analyzed.
Programming Ozobots for teaching astronomy Mariana Balaton, Jorge Cavadas, Paulo Simeão Carvalho, J J G Lima Physics Education, 2021 Experimental teaching is essential for a good understanding of science, especially on Physics. Practical activities play an important role for engaging students with science, mainly when they interact directly with equipment, collect experimental data with computers and/or use interactive software for data analysis. In this work, we present the use of low-cost mini-robots as an ‘object-to-think-with’ for teaching and learning with technology. The activity concerns programming the robots to make them run in circular paths, record videos of their trajectories and analyse them with Tracker Software, to boost the study of Astronomy contents. This kind of practical activity develops multiple skills in students and is usually very well accepted because it involves robots, programming, manipulating technology and for raising topics that are difficult to understand in real-life observations, making them cognitively accessible to the vast majority of students. In this practical activity, students are asked to create the robot programming code and make a video recording (with a smartphone) of the robots’ trajectories, mediated by the teacher, who assists in the construction of the experimental activity and analysis of the data obtained. The results will allow students to understand Kepler’s laws of planetary motion and why some planets seem to have an apparent retrograde motion as seen from the Earth, a problem that arose in IV BC and was only officially solved by the Copernicus heliocentric model, published in 1543, the year of his death.
Accretion in low-mass members of the Orion Nebula Cluster with young transition disks R. M. G. de Albuquerque, J. F. Gameiro, S. H. P. Alencar, J. J. G. Lima, C. Sauty, C. Melo Astronomy and Astrophysics, 2020 Context.Although the Orion Nebula Cluster is one of the most studied clusters in the solar neighborhood, the evolution of the very low-mass members (M* < 0.25 M⊙) has not been fully addressed due to their faintness.Aims.Our goal is to verify if some young and very low-mass objects in the Orion Nebula Cluster show evidence of ongoing accretion using broadband VLT/X-shooter spectra.Methods.For each target, we determined the corresponding stellar parameters, veiling, observed Balmer jump, and accretion rates. Additionally, we searched for the existence of circumstellar disks through available on-line photometry.Results.We detected accretion activity in three young stellar objects in the Orion Nebula Cluster, two of them being in the very low-mass range. We also detected the presence of young transition disks with ages between 1 and 3.5 Myr.
Unveiling YSO dynamics through observations and simulations Memorie Della Societa Astronomica Italiana Journal of the Italian Astronomical Society, 2017
Accretion and jet simulations in YSOs Memorie Della Societa Astronomica Italiana Journal of the Italian Astronomical Society, 2017
Is the efficiency of magnetic braking limited by polar spots? European Space Agency Special Publication ESA SP, 2006
On the problem of magnetic braking J. M. FERREIRA, A. AIBÉO, J. LIMA 2005 Past Meets Present in Astronomy and Astrophysics Proceedings of the 15th Portuguese National Meeting Enaa 2005, 2006
Application of an hybrid MHD wind model with latitudinal dependences to the ULYSSES solar wind data at minimum European Space Agency Special Publication ESA SP, 2005
