Enhanced Space Debris Detection and Monitoring Using a Hybrid Bi-LSTM-CNN and Bayesian Optimization

Space debris detection is important to the integrity of space missions and satellites, especially with the increase in the number of satellites and spacecraft in orbit. This paper addresses this by a new concept innovative approach using hybrid Bi-LSTM-CNN architecture which is optimized using Bayesian optimization. The study presents an analysis approach based on the whole combination of machine learning and deep learning, a high-quality space debris detector, that can identify both the kind of object and the size of its RCS. The new method goes beyond what we have done so far and shows better results on a wide range of evaluation parameters, such as accuracy, precision, memory, and F1 score. Also, the study takes up the pragmatic issue of training time, thereby ensuring performance in real time. Esthetic trials on real datasets confirm the fit of the hybrid model, sensitivity, and efficacy with 99.16% and 99.98% detection and prediction of space debris types, respectively. In summary, this paper makes space debris tracking much more robust and mitigates threats associated with spaceflight and satellite operations, but they can offer a lot of information on threats and mitigation measures. The findings suggest that this hybrid model could be augmented with current space debris tracking systems, to increase their predictive power and operational effectiveness. Received: 2 July 2024 | Revised: 15 August 2024 | Accepted: 6 September 2024 Conflicts of Interest The author declares that she has no conflicts of interest to this work. Data Availability Statement The datasets cited in this manuscript are publicly available and can be accessed from the original sources referenced in the text. The data that support the findings of this study are openly available in Kaggle at https://www.kaggle.com/datasets/kandhalkhandeka/sate llites-and-debris-in-earths-orbit/data. No additional data were generated or analyzed during the current study. Author Contribution Statement Ishaani Priyadarshini: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration, Funding acquisition.