Scalable and robust deep learning models for space debris detection across orbital regimes and observation platforms

Space debris poses a significant threat to space missions, requiring accurate and real-time detection across multiple orbital regimes and observation platforms. This study aims to develop a scalable YOLOv12-PSO framework for accurate space debris detection and to investigate its effectiveness in improving real-time performance, robustness, and detection accuracy across orbital environments. A publicly available space debris image dataset containing synthetic and real images was used to train and evaluate the proposed framework under varying orbital conditions. The proposed system integrates YOLOv12 with Particle Swarm Optimization (PSO) for hyperparameter tuning, while preprocessing steps such as resizing, normalization, and data augmentation are applied to improve robustness. The model is trained and tested on the dataset for object detection, optimization, and performance assessment. Experimental results demonstrate that the optimized model outperforms baseline approaches, achieving Precision (0.89), Recall (0.87), F1-Score (0.88), and mAP (0.78). Comparative evaluation against Baseline YOLOV12, YOLOV12v8m, YOLOV12v8s, YOLOV12v5s, GEO-SPANN, and CNN-based methods confirms improved detection accuracy and real-time inference performance under noisy and complex conditions. This work presents a novel scalable deep learning framework integrating YOLOv12 with Particle Swarm Optimization (PSO), enhancing detection accuracy, efficiency, and robustness, while contributing to advanced space situational awareness systems, autonomous space debris monitoring, and improved orbital debris mitigation and management strategies. The proposed YOLOv12-PSO framework supports autonomous space debris monitoring and provides a policy-relevant tool for space traffic management, contributing to improved orbital safety, collision risk reduction, and long-term sustainability of space mission operations and satellite environment governance.