National infrastructure agencies and state lenders financing major EPC projects—roads, dams, power plants, ports—have no reliable independent window onto what is actually happening on site between contractor progress reports. Headcounts and equipment utilisation are chronically inflated in self-reported data, masking slow mobilisation, labour shortages and idle plant that together signal schedule and cost risk months before a contractor admits the problem. The gap between claimed and actual resourcing is where cost overruns are born.
A constellation of sub-metre optical microsatellites, augmented by synthetic aperture radar for cloud-penetrating revisits, can image each project site daily and feed automated object-detection models that count personnel clusters, classify vehicle types and tally heavy equipment by category—excavators, concrete pumpers, cranes, haul trucks. Change detection between passes reveals mobilisation trends, shift patterns and demobilisation that precedes contract disputes. These are not estimates; they are measurable counts traceable to individual image chips.
For a sovereign infrastructure ministry or a state development bank, owning this capability means independent verification at every disbursement milestone rather than reliance on a commercial vendor whose data access can be restricted, repriced or withheld under export controls. Field inspection teams can be vectored to anomalous sites ahead of scheduled audits, shifting the oversight posture from reactive to anticipatory. The downstream effect is faster corrective action, reduced cost overruns and a contractual audit trail that holds EPC contractors to account.
Frequently asked
Can a satellite genuinely count individual workers on a busy construction site?
Not reliably at typical commercial resolutions. At 0.5 m optical GSD, individual standing workers are marginal targets; the practical approach is to count vehicle and equipment clusters as proxies for crew deployment, then cross-reference with labour manifests. Purpose-built sub-0.3 m or VHSR constellations improve direct headcount accuracy significantly, but remain rare and expensive. A sovereign constellation designed for this mission would be specced at 0.25–0.3 m GSD with a large focal plane.
What's the difference between using a commercial data-as-a-service provider versus owning the constellation?
A commercial DaaS provider (Planet, BlackSky, ICEYE) gives you imagery under licence terms that can be revised, throttled or revoked — and the raw data never fully resides under national jurisdiction. A sovereign constellation means the government controls tasking priority, downlink scheduling, data custody and the inference models. For a country managing $50B+ in EPC obligations, the independent audit capability alone justifies the capital cost of a small microsatellite constellation.
How does SAR complement optical imagery for equipment detection?
SAR penetrates cloud and operates at night, making it indispensable for 24/7 equipment tracking regardless of weather. Metal equipment — cranes, excavators, concrete batching plants — produces strong radar backscatter and can be detected and classified even at 1 m resolution. Optical imagery resolves colour, brand markings and operator presence that SAR cannot. The optimal architecture fuses both: SAR for continuity and optical for high-confidence classification.
How frequently do sites need to be revisited to be useful for contractor performance monitoring?
Once daily is the practical minimum for detecting significant mobilisation changes (equipment arrival/departure, temporary workforce camps appearing). Twice-daily or better is needed to catch deliberate 'show-of-force' gaming, where a contractor floods a site with equipment the night before an inspection and removes it the next morning. A 72-satellite LEO constellation can achieve sub-4-hour median revisit globally, which closes most gaming windows.
Is this technique only useful for large mega-projects, or can it scale down?
The unit economics currently favour projects above roughly 10 km² site area and $500M contract value, where the cost of satellite tasking and analysis is easily offset by even a 1% reduction in overbilling. Smaller projects become viable as constellation capacity increases and per-image costs fall — Spire and Planet have both published pricing trajectories suggesting sub-$5 per km² tasking within this decade.
What legal or contractual hooks allow a project owner to act on satellite data?
The cleanest approach is to embed a 'remote sensing audit' clause in the EPC contract at award, explicitly permitting the owner to use satellite imagery as a trigger for independent resource verification. Without such a clause, satellite data typically functions as internal intelligence that informs a ground inspection request rather than directly triggering a penalty. Several GCC national procurement frameworks have begun incorporating such clauses following high-profile disputes on energy infrastructure projects.
What happens when a contractor operates night shifts to avoid satellite detection?
SAR sensors are unaffected by darkness and can detect active machinery and light vehicle movement at night. Thermal infrared sensors on future platforms can identify heat signatures from operating engines and concrete curing. A well-designed sovereign constellation would carry at minimum a SAR payload or a thermal IR imager alongside visible-band sensors precisely to close the night-shift surveillance gap.
How does a government procure and operate this capability without full aerospace expertise in-house?
The standard model is a prime system integrator (often ESA-affiliated or a tier-1 defence contractor) who builds, launches and commissions the constellation, followed by a knowledge-transfer phase where national staff assume operations. ESA's ECSS standards (ECSS-E-ST-70 series) provide the operational framework. Several nations have followed this path for Earth observation: UAE's Falcon Eye, Saudi Arabia's SAUDI-SAT series, and South Korea's Kompsat programme are direct precedents.