harshinde/spacenet-rio

SpaceNet (Rio de Janeiro) - Building Detection

This dataset contains high-resolution satellite imagery and corresponding building footprint annotations for (Rio de Janeiro) from the SpaceNet Building Detection Challenge. It is designed for training deep learning models for semantic segmentation and building footprint extraction.

Dataset Details

• Total Tiles: 6,940 image tiles
• Imagery Types:
  • 3-band (RGB) Pan-sharpened GeoTIFFs (high spatial resolution)
  • 8-band Multispectral GeoTIFFs
• Annotations: Building footprints in GeoJSON format (vector polygons), easily convertible to binary raster masks (0: background, 1: building).

Uses

Direct Use

This dataset is directly applicable for:

• Semantic segmentation (binary or instance segmentation)
• Building footprint detection
• Testing new earth-observation model architectures (like U-Nets)

Related Model

A fully configured PyTorch U-Net trained on this exact dataset can be found here: harshinde/spacenet-models

Source

This data originates from the SpaceNet Challenge series, which aims to accelerate the development of open source algorithms for automating mapping from commercial satellite imagery.