Cold-rolled metal strip surface quality inspection system
System Features
Implementation of a neural network model for detecting and classifying defects, featuring the ability to retrain the machine vision-based system.
Integration flexibility
User-friendly interface, reduced operator workload.
Centralized management
The system can be implemented as part of the enterprise's overall digital ecosystem.
Use of digital twins
Camera data is used to predict potential defects.
Neural networks
— powerful trainable algorithms for data classification, requiring significant resources but providing high accuracy.
Automation of the quality assessment process
— by combining machine vision and analytics in a single interface.
Import substitution
The project replaces the German technology of IMS and Parsytec companies. Compared to foreign vendors, our system surpasses them in all metrics.
Detection accuracy
The system has a defect detection and classification accuracy of over 90%.
SYSTEM INTERFACE
