GeoFuse

July 6, 2026 · View on GitHub

GeoFuse is a novel cross-modal fusion framework that leverages road map semantics alongside satellite imagery. We enrich satellite representations with spatial geometric cues that remain robust under varying weather conditions by incorporating geo-aligned road map tiles. In particular, we extend popular benchmarks, i.e., University-1652 and DenseUAV, with corresponding road map data, providing structural priors that complement visual features and enhance cross-platform matching.

Extensive experiments validate that GeoFuse consistently surpasses current state-of-the-art methods, achieving approximately +3.46% and +23.18% Recall@1 gains on the University-1652 and DenseUAV benchmarks against diverse weather conditions, respectively. For more details, please refer to our paper: Road Maps as Free Geometric Priors: Weather-Invariant Drone Geo-Localization with GeoFuse

framework

Dataset

Download Road Map

Citation

@article{fang2026road,
  title={Road Maps as Free Geometric Priors: Weather-Invariant Drone Geo-Localization with GeoFuse},
  author={Fang, Y and Wang, T and Zheng, Z},
  journal={arXiv preprint arXiv:2605.14925},
  year={2026}
}

Google Map Link

Qwen-Image-Edit Code

🔗 Ecosystem

Explore our ecosystem for UAV & Spatial Intelligence 🚁

🚁 UAV & Spatial Intelligence

🎓 The University-1652 Family

🎓

University-1652
_{Multi-view Multi-source Benchmark
Ground · Drone · Satellite · ACM MM'20}