DrishX tracks trucks from Sentinel-2

Clever geospatial engineering with real physics behind it, but the value is in trend monitoring more than anything resembling live intelligence.

• –Free satellite data and a local FastAPI/browser UI make the stack unusually accessible for OSINT and logistics work.
• –The project’s own docs note it was trained on German autobahns, so performance will be uneven outside paved, high-contrast, low-cloud corridors.
• –It can estimate volume, speed, and heading, but it cannot identify vehicle type, cargo, or intent at object level.
• –Best fit is corridor-level change detection for trade, supply chain, and regional activity analysis rather than tactical surveillance.
• –The open-source angle matters: it turns an expensive niche capability into something analysts can reproduce and audit.