Rolled Steel Defect Detection
Rolled Steel Defect Detection: enterprise AI for 100% surface inspection of rolled metal
Rolled Steel Defect Detection is an AI-powered visual inspection system for cold-rolled steel. It replaces manual quality control with automated, real-time defect detection over the full metal surface. The product is a standalone AI software product (also available as DSaaS, Data Science as a Service) built on the EYE Platform. It is at TRL 9 (technology readiness level: proven in industrial operation).
What it does
Computer vision and machine learning identify, categorise and map surface defects in real time on cold-rolled, hot-rolled, galvanized and polymer-coated metal. The system connects line-scan cameras, edge computing nodes, backend databases and operator workstations into one pipeline. Defect topography is logged for full traceability.
Key capabilities
100% surface coverage at line speeds of 30 m/s and above
Detection of defects from 0.5 mm² at optical contrast (Weber) ≥ 0.25
Classification of 30+ defect types in three groups: mechanical and rolling flaws, surface and chemical irregularities, edge and structural imperfections
Live defect telemetry and a real-time counter by category for operators
Integration with production line control (ASU TPP), MES and ERP, including identifying which sheet in a batch has a defect for automated cutting
Autonomous retraining on new data (active learning) for new defect types and steel grades
Target industries: rolling and drawing of purchased steel; rolled steel shape manufacturing.
Delivery and licensing: out-licensing for industrial enterprise integration, with milestone-based payments tied to deployment and scaling phases. Available in Enterprise, Enterprise Ultra, Enterprise Ultima and Infrastructure classes, agreed individually with the AI-Systems team.
More details: https://www.loranjacobs.com/my-projects/eye-defect-detection-rolled-steel
