A Real-Time Automated Defect Detection System for Ceramic Pieces Manufacturing Process Based on Computer Vision with...

Area of Science:

• Materials Science
• Computer Science
• Industrial Engineering

Background:

• Automated defect detection is crucial for industrial quality control.
• Specialized techniques are needed for inspecting unique surfaces like ceramics.
• Existing methods often lack precision for ceramic anomaly detection.

Purpose of the Study:

• To propose and develop an advanced defect detection solution for ceramic pieces.
• To implement a computer vision system with deep learning for real-time industrial application.
• To enhance quality control in ceramic manufacturing through automated inspection.

Main Methods:

• Image acquisition and a dedicated labeling platform for dataset creation.
• Image preprocessing techniques tailored for ceramic surface analysis.
• Convolutional Neural Networks (CNNs) for real-time defect identification.

Main Results:

• The system achieved 98.00% accuracy in defect detection.
• An F1-Score of 97.29% demonstrates the system's effectiveness.
• Successful implementation in a real-world industrial setting (tableware manufacturing).

Conclusions:

• The developed automated system significantly improves defect detection on ceramic pieces.
• The computer vision and deep learning approach offers a precise and effective solution for specialized surfaces.
• This technology enhances quality control and efficiency in the ceramic manufacturing industry.