Revisión de la Literatura sobre el Uso del Aprendizaje Profundo Enfocado en Sistemas de Inspección Ópticos Automatizados para la Detección de Defectos Superficiales en el Sector de la Manufactura

The manufacturing sector uses supervised machine learning methodologies to improve inspection processes through machine vision. Automated optical inspection offers efficiency in the inspection process for the detection of defects in the manufacture of various products. This work contributes with the identification of those limitations in data processing based on the defined set of rules and the management of the process domain. A literature review is proposed on the use of deep learning focused on automated optical inspection systems for surface defect detection in the manufacturing sector. The proposed objective is to identify the different architectures oriented on convolutional neural networks applied in optical inspection systems in order to automate the extraction of features or patterns.