GeoAI is a Python package for applying artificial intelligence to geospatial data analysis and visualization.
The package addresses the challenge of integrating machine learning with geographic information by providing a unified framework that handles the full workflow from data acquisition through model inference. It combines popular AI frameworks including PyTorch, Transformers, and PyTorch Segmentation Models with specialized geospatial libraries, allowing users to work with satellite imagery, aerial photographs, and vector data. The tool abstracts complex machine learning workflows through high-level APIs while maintaining flexibility for advanced customization, and it includes automatic GPU acceleration when available.
GeoAI suits researchers and practitioners working with remote sensing data who need to perform classification, detection, or segmentation tasks without building infrastructure from scratch. The package supports multiple geospatial data formats including GeoTIFF, JPEG2000, GeoJSON, Shapefile, and GeoPackage. A distinctive feature is its QGIS integration through a dedicated plugin, which enables users to run AI workflows directly within the QGIS desktop environment without writing code. The tool also provides interactive visualization through Leafmap and MapLibre integration, and includes extensive notebook examples alongside a comprehensive book resource.
The project maintains active development with regular updates to its core modules and examples. The codebase demonstrates consistent refinement of its API design and documentation. Development activity shows ongoing attention to both new feature additions and maintenance of existing functionality across the package's multiple modules.
