Sat2Flow: A Structure-Aware Diffusion Framework for Human Flow Generation from Satellite Imagery
This provides a scalable solution for urban mobility modeling in data-scarce environments, though it appears incremental as it builds on diffusion models with specific adaptations for spatial topology.
The paper tackles the problem of generating Origin-Destination flow matrices for urban mobility analysis by proposing Sat2Flow, a diffusion-based framework that uses only satellite imagery as input, eliminating the need for costly auxiliary features and ensuring structural coherence under regional reindexing. Experimental results show it outperforms existing baselines in numerical accuracy while preserving distributions and spatial structures.
Origin-Destination (OD) flow matrices are essential for urban mobility analysis, underpinning applications in traffic forecasting, infrastructure planning, and policy design. However, existing methods suffer from two critical limitations: (1) reliance on auxiliary features (e.g., Points of Interest, socioeconomic statistics) that are costly to collect and have limited spatial coverage; and (2) sensitivity to spatial topology, where minor index reordering of urban regions (e.g., census tract relabeling) disrupts structural coherence in generated flows. To address these challenges, we propose Sat2Flow, a latent structure-aware diffusion-based framework that generates structurally coherent OD flows using solely satellite imagery as input. Our approach introduces a multi-kernel encoder to capture diverse regional interactions and employs a permutation-aware diffusion process that aligns latent representations across different regional orderings. Through a joint contrastive training objective that bridges satellite-derived features with OD patterns, combined with equivariant diffusion training that enforces structural consistency, Sat2Flow ensures topological robustness under arbitrary regional reindexing. Experimental results on real-world urban datasets demonstrate that Sat2Flow outperforms both physics-based and data-driven baselines in numerical accuracy while preserving empirical distributions and spatial structures under index permutations. Sat2Flow offers a globally scalable solution for OD flow generation in data-scarce urban environments, eliminating region-specific auxiliary data dependencies while maintaining structural invariance for robust mobility modeling.