This is the official repository for "StreetviewLLM: Extracting Geographic Information Using a Chain-of-Thought Multimodal Large Language Model".
StreetViewLLM integrates large language models with multimodal data to enhance geospatial predictions, leveraging street view imagery, geographic data, and retrieval-augmented generation. Tested in seven major cities, it outperforms traditional models in predicting urban indicators including population density, NDVI, and building height.
Before you start, install dependencies:
pip install -r requirements.txtAPIs needed:
MIT License