What resolution do we actually need to detect new buildings versus monitoring city-wide growth trends?

For city-wide growth statistics and national reporting (e.g. SDG 11.3.1), 10–30 m resolution from Sentinel-2 or Landsat is sufficient. For detecting individual structures or informal dwelling units, you need 1–5 m optical or sub-5 m SAR. For enforcement-grade evidence linking a specific plot to an unauthorised structure, sub-metre data is necessary. A sovereign constellation can tier these products from a single set of spacecraft by combining different sensor types.

Can we do this with free Copernicus data instead of building our own satellites?

Copernicus Sentinel data is free and scientifically excellent, but your tasking priorities will always be subordinate to ESA's programme schedule and European policy objectives. You cannot guarantee daily revisit over your capital city during a crisis, you cannot classify data above a certain sensitivity level, and access can be suspended or throttled. A sovereign constellation gives you guaranteed access, sovereign tasking authority, and the ability to withhold or classify data on national-security grounds.

How do we handle the fact that our country has heavy cloud cover for six months of the year?

The answer is SAR. Synthetic Aperture Radar penetrates cloud and operates day and night, making it the default for tropical, equatorial, and monsoon-affected nations. A small constellation of X-band or C-band SAR microsatellites — 6 to 12 spacecraft — can deliver sub-weekly revisit regardless of weather. ESA's ICEYE and Capella Space have demonstrated this architecture commercially; the technology is mature enough to procure domestically under a national space programme.

What is the SDG indicator this application directly supports?

SDG 11.3.1 measures the ratio of land consumption rate to population growth rate in cities, and it is officially computed using built-up area data derived from satellite imagery — specifically the Global Human Settlement Layer produced by the European Commission's Joint Research Centre. A sovereign built-up area detection capability allows your national statistics office to produce this indicator independently rather than relying on an externally produced global dataset that may not reflect local definitions of urban land.

How accurate are current machine-learning classifiers for built-up area detection?

State-of-the-art deep learning models (U-Net and transformer architectures applied to Sentinel-1/2 fusion) routinely achieve overall accuracies above 90% and F1 scores of 0.85–0.92 in well-validated settings, according to published benchmarks in the ISPRS Journal of Photogrammetry and Remote Sensing. Accuracy degrades to 75–85% in heterogeneous peri-urban zones and slums where building materials blend with background. Sovereign operations should budget for a ground-truth validation programme and accept that no classifier eliminates the need for sample-based auditing.

How many satellites do we actually need for a viable national constellation?

For a medium-sized nation (500,000–2,000,000 km²) targeting weekly cloud-free optical coverage and bi-weekly SAR coverage, a constellation of 6 optical microsatellites plus 4 SAR microsatellites in complementary LEO orbital planes is a defensible starting architecture. This provides redundancy, avoids single-point failure, and can be procured and launched in two tranches over three to four years. Organisations such as ESA's Phi-Lab and national space agencies in South Korea (KARI) and the UAE (MBRSC) have demonstrated comparable programmes at similar scale.

What happens to historical baseline data if we switch from a commercial vendor to our own system?

Historical continuity is a genuine risk: if your built-up area time series was built on a vendor's archive, terminating that contract severs the historical record. The mitigation is threefold — negotiate perpetual data rights into any commercial contract from day one; ensure data is stored in open formats (GeoTIFF, cloud-optimised GeoTIFF, or OGC-compliant services) under national custodianship; and overlap commercial and sovereign collection by at least 12 months to validate cross-calibration. CEOS (Committee on Earth Observation Satellites) publishes interoperability guidelines for exactly this continuity problem.

Can satellite-derived built-up area data hold up in a legal or administrative proceeding?

In an increasing number of jurisdictions, yes — but it depends on the evidentiary chain. The imagery must have a verifiable, tamper-evident acquisition record, known geolocation accuracy (ideally better than 5 m CE90), and a documented processing methodology. Sovereign satellite operators have a strong advantage here because they control the full chain of custody. Several court systems — including in India and Brazil — have admitted satellite imagery as supporting evidence in land encroachment cases, but practice varies widely and national legislation governing geospatial evidence is still evolving.

9.2.1Live

Built-Up Area Detection

Dense informal housing spreading across a flat urban fringe with construction underway — Aerial views of city buildings like this one help satellites track urban expansion and guide smarter, more sustainable city planning.

Detecting Urban Expansion from Orbit

1
SAR / Optical Satellite
Multi-temporal imagery captures surface reflectance
2
Ground Reception
Raw scenes downlinked to regional ground station
3
Change Detection Engine
ML classifier maps new impervious surfaces
4
Urban Planner Dashboard
Growth alerts guide zoning and infrastructure decisions

Satellites image cities repeatedly, and algorithms flag newly built areas so planners can respond to rapid urban expansion in near real time.

Page in the build queue

This application has its place in the atlas. The full page — short and long descriptions, what-matters bullets, sovereignty rationale, and the eight-field architecture block — is being written next.

Maturity reading: Operational systems exist with paying customers. Phase 1 built §4.1 (Maritime Intelligence) end-to-end as the depth-and-format reference. Section 4 is the next section to graduate from scaffold to full content; the rest follow on the roadmap.

Related applications

All in urban planning

9.2.2 · Next

Zoning Compliance Monitoring

Bring this into your sovereign programme.

All solutions Speak with engineering