Request access to this annotated dataset

We review requests manually β€” usually within 1-2 business days.

This is a premium annotated dataset. By requesting access, you agree to credit UtilityMeters.ai as the data source in any research, product, or publication that uses it, and confirm you will not resell or redistribute the raw files.

Log in or Sign Up to review the conditions and access this dataset content.

πŸ“Š Annotated Mechanical Utility Meter Dials Dataset

Solve the "half-digit" problem in automated meter reading.

16,324 expertly labeled images of mechanical utility meters, captured in real-world, in-the-wild conditions across water, gas, electricity, and heat meters from 100+ European manufacturers.

πŸ” This is a preview / sample listing. The full dataset, plus a free bonus pre-trained model, is available at utilitymeters.ai.

image

image

image


Dataset Summary

Legacy Automated Meter Reading (AMR) systems rarely fail on clean, static digits β€” they fail on analog wheels caught mid-rotation. This dataset was purpose-built to close that gap, with dedicated annotation classes for rolling-digit transition states in addition to standard digits.

Key features:

🧩 Spinning-drum coverage 10 dedicated rolling-digit classes (0_1 β†’ 9_0) teach models to resolve overlapping intermediate wheel states
🌍 True in-the-wild robustness Low basement lighting, flash glare on glass covers, extreme viewing angles, scratched/cracked/dirty devices
🏭 Cross-brand generalization 100+ European hardware brands prevent overfitting to casing shape or drum color β€” models learn to read digits, not housings
⚑ VRAM-optimized All images standardized to 640px on the longest side for efficient GPU memory usage during training

🎁 Bonus: Free Pre-Trained Model

Every purchase of the full dataset includes a production-ready YOLO model trained on this exact 22-class taxonomy, achieving 94% mAP50-95. Skip months of R&D and deploy straight to production.


Dataset Structure

Downloads last month
169