Publicação
Computer vision-based quality inspection for additive manufacturing
| Resumo: | This document aims to develop a highly adaptable computer vision-based quality inspection tool. A review of some of the latest computer vision algorithms developed to assist in compo- nent inspection is conducted. The recent advancements in artificial intelligence, machine learn- ing, and IoT enable progress toward a more autonomous and error-reduced environment. Ad- ditive manufacturing is a recent technology and, therefore, does not yet provide a completely efficient system that instills confidence in its users. With the goal of avoiding waste of time and materials in additive manufacturing, a plan for developing a generalizable quality inspec- tion system for various environments is presented. The solution is based on a system com- posed of a Raspberry Pi, a 3D printer, and a high-resolution camera for capturing images. The captured images will be evaluated using a pre-designed and tested deep learning model. This project will also have a graphical interface for user communication and feedback for system optimization. With the developed system, it will be possible to streamline the model training time and extrapolate to various manufacturing environments with minimal changes to the core architecture. |
|---|---|
| Autores principais: | Ferreira, André Filipe Bruno |
| Assunto: | Computer vision Additive Manufacturing Artificial Intelligence Quality Inspection Industry 4.0 Industry 5.0 |
| Ano: | 2023 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso aberto |
| Instituição associada: | Universidade Nova de Lisboa |
| Idioma: | inglês |
| Origem: | Repositório Institucional da UNL |
| Resumo: | This document aims to develop a highly adaptable computer vision-based quality inspection tool. A review of some of the latest computer vision algorithms developed to assist in compo- nent inspection is conducted. The recent advancements in artificial intelligence, machine learn- ing, and IoT enable progress toward a more autonomous and error-reduced environment. Ad- ditive manufacturing is a recent technology and, therefore, does not yet provide a completely efficient system that instills confidence in its users. With the goal of avoiding waste of time and materials in additive manufacturing, a plan for developing a generalizable quality inspec- tion system for various environments is presented. The solution is based on a system com- posed of a Raspberry Pi, a 3D printer, and a high-resolution camera for capturing images. The captured images will be evaluated using a pre-designed and tested deep learning model. This project will also have a graphical interface for user communication and feedback for system optimization. With the developed system, it will be possible to streamline the model training time and extrapolate to various manufacturing environments with minimal changes to the core architecture. |
|---|