High-Speed OCR — Curvy Aluminum Cans

Deep Learning OCR reads successfully on softdrink can

Reading printed expiration dates, batch codes, and production IDs on curvy aluminium cans is a classic machine vision challenge. Reflective surfaces scatter light unpredictably, curvature distorts character shapes, and high production speeds of 800+ cans per minute leave no room for missed reads.

Deep learning OCR approach

Our Retina AI OCR engine is trained specifically on the worst-case scenarios: glare, curvature, low-contrast printing, and motion blur. Rather than matching pixel templates like traditional OCR, our neural network learns what each character looks like under all real-world conditions.

Performance metrics

  • Speed: 800+ cans per minute on standard 330ml aluminium
  • Accuracy: 99.8%+ on expiration dates and production codes
  • Surface handling: matte, glossy, reflective, painted, laser-etched
  • Character types: alphanumeric, special characters, multilingual
  • Integration: direct output to PLC, MES, ERP systems

Industries using this

Beverage canners (soft drinks, beer, energy drinks), food canning (pet food, ready meals, preserves), pharmaceutical aluminium packaging, and any production line requiring code verification on reflective or curved surfaces.

Test our OCR on your own product samples within 48 hours.

← Back to Home