Datasets:
image imagewidth (px) 180 800 |
|---|
Raw Mechanical Utility Meter Dial & Serial Number Images (Preview Set – 50 Images)
Dual-Target Real-World Meter Photos for Computer Vision & OCR Training
Billions of legacy mechanical utility meters are still running worldwide, and teaching a model to read spinning analog dials alongside engraved serial numbers requires serious visual diversity. This preview release gives you 50 free sample images pulled from our full 9,299-image mechanical meter collection, so you can evaluate quality, format, and annotation style before committing to the complete dataset.
What's Inside
- 100% real, "in-the-wild" photography — no synthetic renders. Meters captured in basements, narrow corridors, and outdoor enclosures under genuine field conditions.
- Dual-target framing: every image guarantees both the mechanical roller/dial interface and the manufacturer's serial number are fully visible in a single frame.
- Mid-rotation "half-digit" edge cases included, ideal for training models to handle intermediate wheel states with precision.
- Realistic environmental noise: glare, dust, scratched covers, low light, and skewed angles — the kind of variance your model will actually face in production.
- Clean, uniform square crop pipeline (matching the full dataset's 640x640 format) for fast, VRAM-efficient training.
Best Suited For
- Custom labeling and annotation pipelines
- Two-stage detection + OCR architectures (dial localization → digit/serial recognition)
- Automated asset tracking systems linking meter readings to hardware IDs
Access
This preview is meant for evaluation purposes. The complete dataset (9,299 images with full commercial licensing for proprietary ML/AI training) is available at:
👉 Get the full dataset
🚀 Ready to Move Beyond Preview Data?
While our free samples let you validate quality and format, production-grade meter-reading systems need production-grade data. That's where UtilityMeters.ai comes in.
📊 Enterprise-Grade Annotated Datasets
Skip the labeling bottleneck entirely. Our premium collections deliver modern electronic utility meters with complete, ready-to-use manual annotations — hand-verified and delivered in both YOLO (.txt) and PASCAL VOC (.xml) formats, so they drop straight into your existing training pipeline.
- ✅ Fully annotated — no in-house labeling required
- ✅ Dual-format export (YOLO + PASCAL VOC)
- ✅ Sourced from 30+ real-world European manufacturers
- ✅ Commercial licensing for proprietary ML/AI applications
👉 Browse the full annotated dataset catalog
🏢 UtilityMeters.ai Official Collection
Datasets are just the start. Visit our main Hugging Face profile for the complete lineup of open-source models, dataset previews, and demos — from raw field photography to production-ready CV pipelines for automated meter reading (AMR).
- Downloads last month
- 48