OpenSatMap: High-Resolution Satellite Images for Map Construction

3,787 high-resolution satellite images provide fine-grained annotations for diverse geographic locations. Collected from Google Maps at 0.3m and 0.15m per pixel resolutions, the dataset is intended for large-scale map construction. Author z-hb uploaded it to Hugging Face in May 2025.

Use Cases

• Constructing large-scale maps based on high-resolution satellite imagery.
• Training autonomous driving perception models based on fine-grained geographic annotations.
• Performing geospatial analysis based on diverse geographic locations mentioned in the description.

Strengths

• Contains 3,787 high-resolution satellite images.
• Offers fine-grained annotations for detailed analysis.
• Includes imagery at two distinct resolutions: 0.3m/pixel and 0.15m/pixel.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect geographic bias inherent to the source platform.

Provenance

Source

Google Maps

Collection Method

Collected from Google Maps at level 19 and level 20 resolution.

Freshness

Last updated 2025-05-15 08:39:57.

Geography

Diverse geographic locations, covering popular driving datasets.