Ore detection: Datasets & Models

Manually inspecting ore on a conveyor or in a stockpile is slow, inconsistent, and pulls skilled operators away from higher-value tasks, and human judgment drifts across shifts. Automating ore detection with a trained model gives you consistent, round-the-clock analysis of material flow, fragmentation, and grade indicators without stopping the line. For a mining and metals operation, that means catching oversized rock before it jams a crusher, tracking throughput, and building reliable data from imagery you are already capturing. A model like RF-DETR handles the cluttered, overlapping fragments typical of an ore dataset while remaining fast enough for real-time deployment.

Teams get repeatable, automated detection that runs on existing site cameras, drones, and conveyor imaging without new specialized hardware. Detections integrate into dashboards, sorting controls, and quality reporting pipelines, and the model scales from a single crusher to multiple sites as your data grows. Starting from an open ore dataset lets you fine-tune quickly and reach production accuracy on your own material.

1. Upload an image: Drop any photo into the widget above to try ore detection for free, right on this page.
2. Preview instant results: The model detects ore chunk, waste rock, mineral vein, conveyor belt, haul truck within seconds and returns an annotated image with structured predictions, so you can immediately see how well it works. Adjust the detection classes and re-run until the results fit your scenario.
3. Like what you see?: With a free Roboflow account, you can deploy this workflow with the Serverless Cloud API, customize it in your workspace, or train RF-DETR on your own images for production accuracy.
4. Get started for free: No code or training data required to begin: upload, review the results, and make it part of your stack in a single click.