Satellite Image Super Resolution
Deep networks restore detail lost in low-resolution imagery, giving farm, city and infrastructure analysis better data to work with.
Request a product demoThe problem
High-resolution satellite imagery is expensive and limited, and free imagery cannot show small details. Small plots, minor roads and buildings blur together.
Capabilities
- Visible and multispectral bands
- Sharper edges on roads, buildings and plots
- Georeferencing preserved
- Evaluation with metrics such as PSNR and SSIM
- Batch processing for large areas
- GeoTIFF output
Technical specifications are set for each project.
How it works
Ingest
The satellite scene and its spectral bands are loaded.
Prepare
Corrections are applied and the scene is split into tiles.
Reconstruct
A deep network reconstructs the detail in each tile.
Evaluate and export
Results are measured with quantitative metrics and exported as a georeferenced image.
Where it's used
- Monitoring small farms
- Urban change detection
- Infrastructure monitoring
- Better input for detection models
Related services and products
- Photogrammetry and Remote SensingDrone and satellite imagery turned into 3D models, maps and analysis layers.
- Survey and MappingField and aerial surveys, delivered as maps and terrain models with checked accuracy.
- Artificial IntelligenceCustom deep learning models, from raw data to running in your systems.
Common output formats
For sensitive uses, output is checked against reference imagery, because reconstructed detail is a model estimate and not a direct observation.
Have a problem that needs solving? We'll test the solution on real data.
Write a few lines about your project and the Fanrob team will get in touch.
Request a product demo