Technical lecture

The manufacturing industry is setting forth new standards and requirements for quality inspection in order to respond effectively to the market’s demands. In this context, computer vision methods in industrial product surface defect detection have been identified as a promising solution to enhance the inspection processes, reducing the impact of this repetitive task on human operators’ health. However, detecting defects on ceramic tableware poses technical challenges due to the reflective and non-flat surface characteristics of the products, necessitating efforts to ensure detection accuracy. This project proposes the implementation of an automatic inspection system based on the YOLOv8 Deep Learning model, and customized setup for real-time acquisition of ceramic plates’ images. The proposed system detects defects based on their industrial classification, achieving a mean accuracy of 98 %. This prototype can provide comprehensive insights into ongoing activities and product descriptions in real-time, ensuring a dynamic and informed manufacturing environment.