Highways and roads segmentation

The segmentation of highways and roads in satellite images is a fundamental process for applications in the field of geospatial intelligence. By using high-resolution (HR) and very high-resolution (VHR) images with a precision of 1.5 m/pixel or less, the system allows for the precise detection, delineation, and mapping of road networks, improving planning, management, and intervention in both civil and military environments.

This use case focuses on the segmentation of highways and roads and paths in satellite images, enabling the accurate tracing of road networks through advanced Deep Learning algorithmsSpecifically, its goal is to extract and represent roads and paths as lines after processing the resulting pixel masks. líneas tras el procesamiento de las máscaras de píxeles resultado.

The vectorization process generates the following output reports:

• Vector lines of roads:representations of the road networks extracted from the pixel masks.
• Geospatial visualization:maps showing the location and shape of the detected roads overlaid on satellite images.

These results can be exported in standard geospatial formats (such as GeoJSON or Shapefile) for integration into other geographic information systems (GIS). GeoJSON o Shapefilefor integration into other geographic information systems (GIS).