Governments administering rapidly urbanising countries face a fundamental data problem: informal settlements grow faster than any census or field survey can track them. A neighbourhood that housed 40,000 people in last year's municipal plan may house 65,000 today, with new structures appearing on flood plains, steep slopes, or utility easements that no planning authority has sanctioned. Without timely, spatially precise data, budgets for water, sanitation, schools and roads are allocated to a city that no longer exists.
A constellation of sub-5m multispectral microsatellites, backed by an X-band SAR layer for cloud-penetrating revisits in tropical wet seasons, delivers a wall-to-wall change-detection product every 7–14 days. Machine-learning classifiers trained on local building morphology distinguish formal from informal structures, flag new footprints, and estimate roof-material proxies for structural vulnerability. The whole pipeline runs on sovereign compute, so the underlying settlement geometry never transits a foreign cloud.
The operational outcome is a live urban cadastre that planners, utilities and social-welfare agencies share from a single source of truth. Ward-level growth alerts trigger field verification teams within days rather than years. Over a five-year cycle the government accumulates a time-series baseline that is already reshaping how the World Bank and UN-Habitat negotiate slum-upgrading loans—because the data are owned, controlled and auditable by the state that must act on them.
Frequently asked
Why should a government own a slum-mapping satellite capability rather than simply buying imagery from Planet or Maxar?
Sovereign ownership eliminates the risk that a foreign commercial vendor reprices, restricts or discontinues service during a political crisis or budget negotiation. Informal settlement data is politically sensitive — it underpins land-reform policy, eviction decisions and donor funding — so control of the collection schedule, data classification and release authority must sit with the national government. Buying imagery as a service cedes that control permanently.
What resolution do satellites need to usefully map individual structures in a dense informal settlement?
Structure-level delineation requires ground sample distances of 0.3 m to 0.5 m, achievable by current VHR optical microsatellites such as Planet's SkySat or Airbus's Pléiades Neo. Neighbourhood-level growth tracking can be performed at 3–5 m resolution using constellations like Sentinel-2 or a national medium-resolution system, which dramatically reduces per-image cost and enables daily revisit cadences.
How frequently must imagery be collected to detect meaningful settlement growth?
In fast-growing cities, meaningful structural change — new rooflines, track extension, in-fill densification — can accumulate over 30–90 days. Quarterly collections are the practical minimum for policy-relevant trend lines; monthly collections are preferred for early-warning applications. A LEO microsatellite constellation of 12–36 satellites achieves daily revisit, well exceeding that requirement.
Can satellite-derived slum maps be used as legal evidence for tenure recognition?
Satellite mapping can establish proof of occupancy duration and settlement boundary — evidence cited in several World Bank-financed upgrading projects and in UN-Habitat's global slum-upgrading toolkit. However, courts and land registries in most jurisdictions require corroborating field surveys and administrative records. Satellite data functions as a powerful first-pass filter and audit layer, not a standalone legal instrument.
What happens to the accuracy of AI-based classifiers when applied to a city the model was not trained on?
Generalisability is the central technical challenge: a model trained on Nairobi's corrugated-iron roofscapes underperforms on Dhaka's bamboo-and-tarpaulin settlements or Lagos's concrete-block densification. Transfer learning and fine-tuning on local ground-truth samples recovers most accuracy, but this requires each implementing agency to maintain labelled local training sets and periodic model retraining — a capacity that many national mapping agencies do not yet have.
Which international organisations publish standards or methodologies that a national programme should align with?
UN-Habitat publishes the global slum definition and enumeration methodology used in SDG Indicator 11.1.1 monitoring. The World Bank's Open Data Platform provides upgrading project benchmarks. ISO/TC 211 standards (ISO 19115, ISO 19157) govern geospatial metadata and data quality reporting, while OGC interoperability standards ensure outputs can be ingested by national spatial data infrastructures and donor monitoring systems.
How does SAR complement optical imagery for slum mapping?
Synthetic aperture radar penetrates cloud cover and operates at night, making it indispensable in tropical and monsoon climates. SAR coherence change detection can flag new construction with high sensitivity even at 6–10 m resolution. Fusing SAR (e.g. Sentinel-1, ICEYE, Capella Space) with optical data improves classifier robustness and reduces data-gap periods caused by weather, though SAR image interpretation requires specialist training.
What is the link between slum mapping and SDG progress reporting?
SDG Target 11.1 requires countries to report the proportion of the urban population living in slums, squatter settlements or inadequate housing (Indicator 11.1.1). Satellite-derived settlement boundaries, when fused with census denominators, provide the most scalable and reproducible method for meeting the UN-Habitat reporting methodology, especially in countries with irregular census cycles. Nations that own their mapping capability can update this indicator annually rather than waiting a decade for a new census.