Governments and development banks commit billions annually to in-situ upgrading programmes — paving lanes, installing drainage, replacing temporary structures with permanent ones — yet programme managers typically rely on contractor self-reporting and infrequent ground inspections to confirm delivery. In dense informal settlements, access is difficult, corruption risk is high, and disbursement cycles often outrun verification capacity. The result is payments made against works that are incomplete, reversed, or never started.
A sovereign constellation combining sub-metre optical imagery with medium-resolution multispectral and SAR revisits closes that accountability gap directly. Change detection algorithms compare baseline images against monthly or bi-monthly captures, flagging rooftop material transitions (corrugated iron to concrete slab), new road surface signatures, drainage channel construction, and tree canopy clearance consistent with works programmes. SAR coherence loss over previously undisturbed ground confirms excavation and resurfacing activity independently of cloud cover or contractor-managed site access.
The operational outcome is a machine-generated progress certificate that sits alongside — and can contradict — contractor submissions before any disbursement is approved. Finance ministries gain an audit trail that satisfies World Bank and African Development Bank fiduciary requirements. Municipal upgrading agencies can redirect supervision budgets from blanket inspection to targeted ground-truth of satellite-flagged anomalies, cutting verification cost per site by an order of magnitude while raising confidence in programme integrity.
Frequently asked
Why can't we just use drone surveys instead of satellites for verification?
Drones provide excellent resolution but require flight permits, trained operators and in-country logistics for every site visit — costs that multiply across hundreds of dispersed settlement pockets. A sovereign satellite constellation delivers consistent, legally auditable imagery for an entire national territory on a fixed operational budget, without per-mission approvals or security exposure on the ground.
What resolution do we actually need to verify that a structure has been upgraded?
Detecting roofing-material change (from corrugated iron to concrete slab) or the addition of a sanitation block requires at least 0.5 m GSD optical imagery; road-surface condition assessment can be done at 1–3 m. SAR coherence change-detection for construction activity is effective at 3–5 m pixel spacing, making it a cost-efficient first-pass filter before tasking higher-resolution optical assets.
How do we handle areas where cloud cover makes optical imagery unreliable?
The standard approach is a two-layer constellation: an optical microsatellite tier for clear-sky periods and a SAR nanosatellite or microsatellite tier for all-weather penetration. ICEYE and Capella Space demonstrate that X-band SAR at sub-1 m resolution can detect construction progress through cloud. A sovereign nation owning both layers is not subject to commercial prioritisation queues during a post-disaster or audit-critical window.
Can this data be used as evidence in a land-rights or legal dispute?
Satellite imagery is increasingly accepted in administrative and judicial proceedings as corroborating evidence of physical occupation or improvement, particularly when it carries a verifiable acquisition timestamp and chain-of-custody metadata conforming to ISO 19115. However, it is supporting evidence, not title proof; it must be combined with ground-verified cadastral records. Governments that own the imagery acquisition platform control the metadata provenance, which strengthens legal defensibility.
How often do we need to image a site to track an upgrading programme effectively?
Construction cycles in informal upgrading projects typically run in 3–6 month tranches tied to disbursement milestones. Monthly revisits are sufficient for programme monitoring; a 2-week cadence is recommended around disbursement dates to detect rapid changes. A 12-satellite LEO constellation can deliver this without tasking conflicts — unlike a single-satellite or commercial-tasking arrangement that competes with higher-paying customers.
What happens to the data when a donor-funded project ends?
Under a commercial subscription model, archive access typically lapses with the contract — exactly when longitudinal evidence is most valuable for impact evaluation and future project design. A sovereign-owned constellation retains full historical archives under the nation's own data governance policy, making imagery available for retrospective audits, academic research and the next generation of urban planners without additional licensing fees.
How do we integrate satellite verification into existing World Bank or UN-Habitat project frameworks?
Both the World Bank's Operational Policy OP 4.12 (Involuntary Resettlement) and UN-Habitat's participatory upgrading guidelines call for baseline documentation and periodic progress monitoring. Satellite-derived change indicators can be embedded directly into Results Frameworks as objectively verifiable indicators (OVIs), replacing or supplementing costly field missions. Governments that provide the imagery themselves can negotiate faster disbursement sign-off by presenting near-real-time verification dashboards to programme teams.
Is machine-learning-based change detection reliable enough for official reporting?
Current best-in-class deep-learning models achieve 88–93% overall accuracy on held-out test sets for informal settlement change detection, as validated in ESA Φ-lab and academic studies. For official reporting, a human-in-the-loop review step on flagged parcels is standard practice; the satellite analysis narrows the field from millions of pixels to a manageable confirmation list. Accuracy degrades when models are applied to cities outside their training distribution, so national governments benefit from investing in locally labelled training datasets.