Utilizando Redes Neurais Convolucionais para Automatizar a Detecção de Defeitos Físicos em PCBs na Indústria 4.0
Keywords:
Convolutional Neural Networks, defect detection, memory modules, VGG19, MobileNet, InceptionV3, accuracy, Industry 4.0
Abstract
Printed circuit boards (PCBs) are becoming increasingly complex and miniaturized, making the detection of physical defects more challenging. Convolutional neural networks (CNNs) can be used to detect complex patterns in images, and this study proposes the use of CNNs for the detection of defects in PCBs. Three canonical CNN architectures were evaluated: VGG19, MobileNet, and InceptionV3. MobileNet showed the best performance, with an accuracy of approximately 98.12%. This study provides a solid foundation for the development of automated defect detection systems for the information technology industry.
Published
2023-10-18
Section
Articles