Multimodal-Multiresolution Foundation Model for Lunar Remote Sensing

The paper introduces a multimodal foundation model for lunar remote sensing, trained from scratch on SomBench—a dataset of nearly two million co‑registered tile bundles across 11 modalities at 1 m and 100 m resolutions. The model extends the TerraMind masked‑token architecture with lunar‑specific features such as explicit acquisition geometry and joint training of two spatial scales, and employs FlexiViT patch embeddings for adaptable patch sizes. Evaluation on crater detection, irregular mare patch segmentation, and polar ice prospectivity regression shows that the pretrained model matches or surpasses ImageNet‑pretrained baselines, with notable label efficiency and effective adaptation via LoRA.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv AI.

• rag
• computer-vision
• fine-tuning
• multimodal
• benchmarks