MineWatch

MineWatch is an AI-powered satellite surveillance platform developed by Crust Solutions Inc. to detect, map and monitor illegal artisanal and small-scale gold mining — commonly known as galamsey — across Ghana's forest regions. The platform combines cutting-edge satellite imagery analysis with real-time intelligence dashboards to give law enforcement agencies, environmental regulators, government bodies and the general public a continuously updated picture of where illegal mining is occurring, how it is spreading, and where intervention is most urgently needed.

Our Mission

Ghana loses an estimated $2 billion annually to illegal mining. The ecological damage — poisoned rivers, destroyed forest reserves, degraded farmland — threatens the livelihoods of millions of Ghanaians. MineWatch exists to give environmental defenders the intelligence they need to act faster, smarter and more effectively against galamsey.

Coverage

MineWatch currently monitors Ghana's six primary mining regions, which account for the vast majority of illegal mining activity in the country:

• Western Region — covering the Tarkwa, Prestea and Bogoso mining belt, one of the most heavily affected areas in Ghana
• Western North Region — covering the Bibiani-Anhwiaso-Bekwai area and surrounding forest zones
• Eastern Region — covering the Kibi, Akyem and Birim River areas
• Ashanti Region — covering the Obuasi area and the forest reserves surrounding Kumasi
• Central Region — covering the Dunkwa and Assin areas
• Ahafo Region — covering the Goaso and Bechem forest areas

MineWatch is actively expanding its coverage. Additional regions across Ghana will be added to the monitoring system in future updates, with the goal of achieving full national coverage across all sixteen regions of Ghana.

How It Works

Every 5 days, the MineWatch detection pipeline automatically retrieves and analyses Sentinel-2 multispectral satellite imagery from the European Space Agency's Copernicus programme. The pipeline compares current vegetation levels against five historical baseline periods — spanning 1 month, 4 months, 6 months, 1 year and 2 years — to identify areas where significant forest loss has occurred. Detections are classified as New Detections, where fresh mining activity has recently appeared, or Previously Detected, where established ongoing mining operations are identified. Nine licensed industrial mining operations are automatically excluded from detection to eliminate false positives from legitimate mining activity.

Reducing False Positives

Generating accurate detections from satellite imagery in Ghana's complex landscape is a significant technical challenge. MineWatch employs multiple layers of filtering to minimise false positives:

• Forest baseline requirement — detections only occur in areas that were forested in the year 2000, automatically excluding towns, villages and farmland
• Settlement exclusion — multiple global settlement datasets are combined to mask out known urban and rural settlements across Ghana
• City and town buffers — over 80 named cities, towns and mining communities have been individually mapped with exclusion buffers
• Forest zone restriction — detections are restricted to Ghana's forest belt, excluding savannah and transitional vegetation zones in the north
• Water body exclusion — permanent water bodies and a 500 metre buffer around the Volta Lake shoreline are excluded
• Minimum area threshold — only clearings of 2 hectares or larger are flagged, filtering out small-scale vegetation disturbances

Data Sources

• Satellite imagery: Copernicus Sentinel-2 via Google Earth Engine
• Forest baseline: Hansen Global Forest Change 2024 (University of Maryland)
• Settlement data: GHSL, ESA WorldCover, Google Open Buildings
• Spatial resolution: 10 to 30 metres per pixel
• Detection threshold: Minimum 2 hectares

MineWatch v1.0 — April 2026 — Developed by Crust Solutions Inc.