PhySwin: An Efficient and Physically-Informed Foundation Model for Multispectral Earth Observation

Chong Tang (University College London) · Joseph Powell (University of Manchester) · Dirk Koch (Ruprecht-Karls-Universität Heidelberg) · Robert Mullins (University of Cambridge) · Alex Weddell (University of Southampton) · Jagmohan Chauhan (University College London, University of London)

computational complexityearth observation imageryefficiency constraintsinference latencylow-flopmixmae formulationmultispectral dataphysical priorsphysics-informed pretrainingphyswinradiometric constraintsremote sensing foundation modelssota resultsswinv2token-efficient spectral embeddinguniversal representations

Recent progress on Remote Sensing Foundation Models (RSFMs) aims toward universal representations for Earth observation imagery. However, current efforts often scale up in size significantly without addressing efficiency constraints critical for real-world applications (e.g., onboard processing, rapid disaster response) or treat multispectral (MS) data as generic imagery, overlooking valuable physical priors. We introduce PhySwin, a foundation model for MS data that integrates physical priors with computational efficiency. PhySwin combines three innovations: (i) physics-informed pretraining objectives leveraging radiometric constraints to enhance feature learning; (ii) an efficient MixMAE formulation tailored to SwinV2 for low-FLOP, scalable pretraining; and (iii) token-efficient spectral embedding to retain spectral detail without increasing token counts. Pretrained on over 1M Sentinel-2 tiles, PhySwin achieves SOTA results (+1.32\% mIoU segmentation, +0.80\% F1 change detection) while reducing inference latency by up to 14.4$\times$ and computational complexity by up to 43.6$\times$ compared to ViT-based RSFMs.