GEAR Lab Research on Satellite-to-Street View Synthesis Presented at IGARSS 2026 🎉

GEAR Lab is pleased to share our research presentation at the IEEE International Geoscience and Remote Sensing Symposium (IGARSS 2026) held in Washington, D.C. The work is co-authored by Yifan Yang, Dr. Lei Zou, and Dr. Wendy Jepson.

The presentation, titled “Satellite-to-Street: Synthesizing Post-Disaster Views from Satellite Imagery via Generative Vision Models,” explores cross-view image synthesis to bridge the gap between overhead satellite observations and street-level disaster perspectives. Assessing ground-level structural damage immediately after extreme events is frequently hindered by physical access constraints and data latency. To address this bottleneck, the framework leverages generative vision models to synthesize realistic ground-level post-disaster views directly from overhead satellite imagery, enhancing rapid situational awareness and damage intelligence.

This work demonstrates how generative AI and remote sensing can be combined to enrich disaster intelligence workflows when on-the-ground observations are delayed or unavailable. By reconstructing ground-level visual context, the proposed pipeline provides emergency managers, planners, and first responders with actionable, hyperlocal spatial insights for early-stage response and post-disaster assessment.

We are also grateful to receive a $500 IEEE GRSS Travel Grant to support presenting this research at the conference. This study was supported by the Texas A&M University Environment and Sustainability Initiative (ESI) through the Environment and Sustainability Graduate Fellow Award. Special thanks to Dr. Wendy Jepson and the ESI community for their continuous support.

Congratulations to the research team on this contribution to GeoAI and remote sensing!

Citation

Yang, Y., Zou, L., & Jepson, W. (2026). Satellite-to-Street: Synthesizing Post-Disaster Views from Satellite Imagery via Generative Vision Models. arXiv preprint arXiv:2603.20697.