Please use this identifier to cite or link to this item: http://hdl.handle.net/10609/150502
Title: Detecció automàtica de defectes de soldadura mitjançant l’anàlisi de visual transformers i tècniques de time series imaging
Author: Velasco Gallego, Christian
Tutor: Isern, David  
Others: Sánchez Castaño, Friman
Abstract: Numerous defects, such as hull damage, engine failures, and equipment malfunctions, can occur in maritime operations, which may affect the safety and reliability of ships. Accordingly, the detection and classification of defects in ships is of paramount importance to guarantee its adequate functioning. This paper introduces a new time series imaging approach for defect identification by combining distinct time series imaging approaches with a vision transformer. Specifically, the time series imaging methods Gramian Angular Field (GAF), Recurrence Plot (RP), and Markov Transition Field (MTF) are analysed in this study. A case study on metal arc welding is also presented to highlight the performance of the proposed methodology and assess the feasibility of implementing this type of methods for defect identification. The results of this case study indicate that Gramian Difference Angular Field is the most feasible encoding method for the task defined, as this method achieved an accuracy of over 70%.
Keywords: identificació de defectes
aprenentatge profund
codificació de sèries temporals en imatges
Document type: info:eu-repo/semantics/bachelorThesis
Issue Date: Jun-2024
Publication license: http://creativecommons.org/licenses/by-nc/3.0/es/  
Appears in Collections:Bachelor thesis, research projects, etc.

Files in This Item:
File Description SizeFormat 
cvelascogTFG0624memoria.pdfMemòria del TFG891,98 kBAdobe PDFThumbnail
View/Open
cvelascogTFG0624presentacio.pdfPresentació del TFG854 kBAdobe PDFThumbnail
View/Open
Share:
Export:
View statistics

This item is licensed under aCreative Commons License Creative Commons