Acceso usuarios
Luiz Carlos da Silva Garcia Junior, Cláudio Henrique Albuquerque Rodrigues, Isomar Lima da Silva
The growing demand for operational efficiency and quality control in Industry 4.0 has driven the adoption of intelligent computer vision systems. This study aims to develop and optimize advanced computer vision algorithms for the detection and classification of cosmetic defects, defined as visual imperfections that affect the aesthetic quality of products. The methodology combines convolutional neural networks, deep learning models, and advanced image processing techniques, fully integrated with industrial sensors and automation systems. The algorithm, implemented in a prototype, was validated in a real manufacturing environment, achieving an accuracy of 98%, with a 90% reduction in false negatives, an 85% reduction in false positives, and a decrease in average inspection time from 5 seconds to 0.3 seconds per unit. These results demonstrate that the adoption of intelligent computer vision significantly enhances quality control accuracy while improving productivity and reliability in industrial processes, reinforcing its applicability in the digital transformation of manufacturing.
