--- pretty_name: AstroDim license: cc-by-4.0 task_categories: - object-detection tags: - image - astronomy - dim-moving-object-detection - image-sequence - spatiotemporal-modeling - point-annotations size_categories: - n<1K --- # AstroDim AstroDim is a real-world astronomical image-sequence dataset for dim moving object detection, introduced in **FalconDet: A trajectory-aware spatiotemporal framework for dim moving object detection**. It provides high-resolution (4096 × 4096) observations for studying the detection of faint moving targets using spatial and temporal information. - **Paper:** [FalconDet — Pattern Recognition](https://www.sciencedirect.com/science/article/abs/pii/S0031320326020017) - **Code Repository:** [CastleChen339/FalconDet](https://github.com/CastleChen339/FalconDet) Citation If you use AstroDim in your research, please cite: ```bibtex @article{chen2026falcondet, title={FalconDet: A trajectory-aware spatiotemporal framework for dim moving object detection}, author={Chen, Jiuchen and Yan, Xinyu and Xu, Qizhi and Han, Xiaolin}, journal={Pattern Recognition}, pages={115037}, year={2026}, publisher={Elsevier} } ```