Recent advancements in space exploration technology have significantly increased the number of diverse satellites in orbit. This surge in space-related information has posed considerable challenges in developing space target surveillance and situational awareness systems. However, existing detection algorithms face obstacles such as complex space backgrounds, varying illumination conditions, and diverse target sizes. To address these challenges, we propose an innovative end-to-end Attention-Guided Encoder DETR (AgeDETR) model, since artificial intelligence technology has progressed swiftly in recent years. Specifically, AgeDETR integrates Efficient Multi-Scale Attention (EMA) Enhanced FasterNet block (EF-Block) within a ResNet18 (EF-ResNet18) backbone. This integration enhances feature extraction and computational efficiency, providing a robust foundation for accurately identifying space targets. Additionally, we introduce the Attention-Guided Feature Enhancement (AGFE) module, which leverages self-attention and channel attention mechanisms to effectively extract and reinforce salient target features. Furthermore, the Attention-Guided Feature Fusion (AGFF) module optimizes multi-scale feature integration and produces highly expressive feature representations, which significantly improves recognition accuracy. The proposed AgeDETR framework achieves outstanding performance metrics, i.e., $97.9\%$ in ${mAP}_{0.5}$ and $85.2\%$ in $mA{P}_{0.5:0.95}$, on the SPARK2022 dataset, outperforming existing detectors and demonstrating superior performance in space target detection.

Small target detection has become an important part of space exploration missions. The existence of weak illumination and interference from the background of star charts in deep and distant space has brought great challenges to space target detection. In addition, the distance of space targets is usually far, so most of them are small targets in the image, and the detection of small targets is also very difficult. To solve the above problems, we propose a multi-scale feature fusion pyramid network. First, we propose the CST module of a CNN fused with Swin Transformer as the feature extraction module of the feature pyramid network to enhance the extraction of target features. Then, we improve the SE attention mechanism and construct the CSE module to find the attention region in the dense star map background. Finally, we introduce improved spatial pyramid pooling to fuse more features to increase the sensory field to obtain multi-scale object information and improve detection performance for small targets. We provide two versions and conducted a detailed ablation study to empirically validate the effectiveness and efficiency of the design of each component in our network architecture. The experimental results show that our network improved in performance compared to the existing feature pyramid.