Illegal mining — whether artisanal gold panning in protected forest or unlicensed industrial excavation — costs resource-rich nations billions in lost royalties, accelerates deforestation, poisons river systems with mercury, and funds organised crime. Ground enforcement alone cannot cover the scale: a single Amazonian nation may have mining concession boundaries spanning hundreds of thousands of square kilometres, much of it roadless and actively hostile. Satellite tasking is the only practical way to generate persistent, evidence-grade intelligence across that geography.
A sovereign constellation combining multispectral imaging with SAR closes both the cloud-cover gap and the camouflage problem. Multispectral bands detect the characteristic spectral signature of disturbed laterite soil, deforested clearings and mercury-contaminated sediment plumes in watercourses. SAR penetrates cloud year-round and reveals freshly cleared ground through canopy where optical misses it. RF survey payloads can passively detect the VHF and UHF communications traffic associated with mining camps. Together they produce a change-detection cadence tight enough to catch a new illegal pit within days of it opening.
The operational output is a timestamped, georeferenced evidence package — before-and-after imagery, SAR coherence change maps, watercourse turbidity anomalies — that environment and mining regulators can hand directly to prosecutors or use to direct enforcement aircraft. Crucially, a sovereign pipeline ensures that tipping intelligence about politically connected operators is not first routed through a foreign commercial vendor whose government may have its own interests in the concession.
Frequently asked
Why can't a government just buy this data from Planet or ICEYE instead of building its own satellite?
Purchasing archive access gives you imagery, but not control over tasking priority, data retention, or revisit scheduling. An illegal mining emergency — a sudden expansion into a protected watershed — may not align with a vendor's standard collection plan. Owning the sensor means the government decides when, where, and how often to look, and it retains the raw data under national sovereignty rather than subject to a foreign vendor's terms of service or export licence.
What sensor combination works best for illegal mining detection?
Multispectral optical imagery (3–5 m resolution) is the baseline for land-cover change and sediment-plume detection in waterways. SAR (X- or C-band, 1–3 m) provides all-weather capability and is essential in persistently cloudy tropical zones. Combining both in a heterogeneous constellation — with optical for classification and SAR for temporal consistency — yields the most robust change-detection pipeline. Hyperspectral is a useful addition for tailings characterisation and mercury-contamination proxies.
How do you distinguish illegal mining from licensed artisanal or small-scale mining (ASM) using satellite data alone?
You cannot make a legal determination from orbit. The satellite layer identifies anomalies: new clearance events, turbidity spikes, equipment signatures, or excavation patterns. Those anomalies are then cross-referenced against a government cadastral database of licensed operations. The delta — change events outside licenced boundaries — feeds the enforcement queue. Maintaining an authoritative, up-to-date mining-licence cadastre is therefore as important as the satellite capability itself.
What resolution is needed to detect small-scale dredging operations on rivers?
Dredge barges typically measure 15–40 m in length; 3–5 m optical or SAR imagery is sufficient for detection and rough classification. Sub-2 m imagery enables counting of individual vessels and resolving equipment type. For sediment-plume mapping, medium-resolution multispectral (10–30 m, e.g. Sentinel-2 class) captures the downstream turbidity signature even when the dredge itself is under canopy cover.
Can machine-learning change detection keep pace with the scale of illegal mining?
Yes — at continental scale, ML pipelines (convolutional neural networks trained on labelled mining-site imagery) now outperform manual interpretation for both speed and consistency, achieving detection F1 scores above 0.85 in published benchmarks. The real bottleneck is alert routing and enforcement capacity, not detection speed. Governments should invest in analyst workflows and inter-agency data-sharing protocols in parallel with the ML pipeline.
How often does the satellite need to revisit the same area to be operationally useful?
Enforcement agencies generally regard 24–48 hour revisit as the minimum for generating actionable intelligence; illegal operations can establish meaningful surface disturbance within 48 hours of starting work. A six-satellite LEO microsatellite constellation in a sun-synchronous or inclined orbit can achieve sub-12-hour revisit at tropical latitudes. Critically, consistent, predictable revisit matters more than occasional very high-frequency tasking.
Is international data-sharing obligatory, and does it constrain what a sovereign operator can do with its own imagery?
No treaty mandates illegal-mining satellite data sharing, though the Minamata Convention (Decision MC-4/8) encourages member states to share monitoring data on mercury-related ASGM. Bilateral and regional arrangements (e.g. Amazon Cooperation Treaty Organization members sharing INPE data) are voluntary. A nation owning its own satellite is free to share selectively, protect sensitive operational images from enforcement raids, and withhold data that could compromise ongoing investigations — options unavailable when relying on commercial providers governed by their home-country export rules.
What is the realistic capital cost of a sovereign illegal-mining monitoring constellation?
A four-to-six satellite LEO microsatellite constellation offering dual optical/SAR capability, including ground segment, data processing infrastructure, and five-year operations, falls in the $120–280M range depending on procurement model and launch costs. Set against $12.9B in annual fiscal losses to artisanal mining tax evasion in sub-Saharan Africa alone (World Bank, 2022), the break-even period for a mid-sized resource-rich nation is well under two years if enforcement conversion rates are modest.