Publication: Deep Learning para la detección de defectos de fabricación industrial en sensores de aparcamiento
Loading...
Official URL
Full text at PDC
Publication Date
2022-09
Authors
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
En el presente Trabajo de Fin de Máster se presentan una serie de arquitecturas de redes neuronales con el objetivo de dar solución al problema de detectar defectos de fabricación en piezas industriales, tales como sensores de aparcamiento. De esta manera, a lo largo del trabajo se introducirá el estado del arte de la visión artificial actual acompañándolo con las correspondientes explicaciones teóricas acerca de los algoritmos usados dentro del contexto del Deep Learning. A continuación, se hará hincapié en el recorrido que se ha seguido durante el ciclo de vida entero del proyecto para, posteriormente, exponer los resultados obtenidos como consecuencia de probar los algoritmos introducidos sobre el conjunto de datos inicial, el cual se habrá presentado con anterioridad, y terminar así extrayendo unas conclusiones finales de estos.
In this Master’s thesis a series of neural network architectures are presented with the aim of solving the problem of detecting manufacturing defects in industrial parts, such as parking sensors. Thus, throughout the work, the state of the art of current computer vision will be introduced along with the corresponding theoretical explanations about the algorithms used in the context of Deep Learning. Then, we will emphasize the path that has been followed during the entire life cycle of the project to, subsequently, expose the results obtained as a result of testing the algorithms introduced on the initial data set, which will have been presented previously, and end up drawing some final conclusions from them.
In this Master’s thesis a series of neural network architectures are presented with the aim of solving the problem of detecting manufacturing defects in industrial parts, such as parking sensors. Thus, throughout the work, the state of the art of current computer vision will be introduced along with the corresponding theoretical explanations about the algorithms used in the context of Deep Learning. Then, we will emphasize the path that has been followed during the entire life cycle of the project to, subsequently, expose the results obtained as a result of testing the algorithms introduced on the initial data set, which will have been presented previously, and end up drawing some final conclusions from them.
Description
Trabajo de Fin de Master en Internet de las Cosas, Facultad de Informática UCM, Departamento de Ingeniería del Software e Inteligencia Artificial, Curso 2021/2022.