The Cartographic change detection is used to identify and monitor transformations in the terrain, such as new constructions or modifications in infrastructure. This use case relies on high-resolution (HR) and very high-resolution (VHR) satellite images that allow for detailed observation of urban changes between two images.
This use case implements a model that combines elements of Siamese networks, CNNs, and Transformersto identify changes in the urban environment through satellite images. By using images of the same region captured on different dates, the model visually compares the data to detect constructions or modifications in the terrain. The model evaluates two images (an initial one and a comparison one) and determines the areas with modifications, thus enabling precise monitoring of changes.
The cartographic change detection process generates the following types of reports:
- Geospatial visualization:graphical representation of areas where modifications in buildings or infrastructure have been identified, allowing for a graphical comparison of the two analyzed images.
- Exportable geospatial layers with changes:ready to be used in GIS systems (e.g., in GeoJSON or Shapefile formats), allowing for additional visualization and analysis on geospatial intelligence platforms.
Practical Applications
Cartographic change detection is useful in multiple practical applications, including:
- Urban growth monitoring:facilitates the analysis of urban expansion, allowing for the identification of new buildings and expansion areas in cities and towns.
- Map planning and updating:helps keep maps updated in real-time, quickly identifying new constructions or infrastructure modifications.
- Security and defense:enables constant monitoring of changes in strategically important areas, providing critical information in the context of military intelligence and planning.
- Disaster management:quickly identifies structural damages and allows for the assessment of the extent of changes in areas affected by natural disasters or conflicts.

