Please use this identifier to cite or link to this item: http://hdl.handle.net/10609/151951
Title: Detección y análisis de huellas de tortugas en imágenes utilizando Deep Learning
Author: Romero Matarin, Carlos  
Tutor: Bas Pujols, Bernat
Others: Fernandez Blanco, Pablo  
Abstract: This study develops an automated system based on advanced deep learning models, such as YOLOv8 and U-Net, aimed at identifying and segmenting turtle tracks in images of their natural environment. The purpose is to enhance conservation strategies through technological tools that improve the accuracy and efficiency of monitoring. YOLOv8 stands out for its ability to perform rapid and precise real-time detection, making it ideal for dynamic contexts. U-Net's functionality is enhanced by meticulous segmentations, simplifying the granular analysis of occupied areas. The integration of both models enables a hybrid strategy that balances speed and precision, optimizing identification and analysis processes. The proposed system replaces conventional manual methods, resulting in cost reductions and increased uniformity in results. Additionally, technical challenges such as variable lighting and complex backgrounds are addressed through advanced preprocessing techniques, including contrast enhancement and data augmentation. Findings show that YOLOv8 and U-Net are effective in real-world contexts, facilitating decision-making based on consistent data. This solution not only optimizes conservation initiatives but also sets a precedent for the application of artificial intelligence in environmental monitoring. The system supports the protection of endangered species and promotes the sustainable use of marine ecosystems, demonstrating the potential of deep learning to tackle complex issues in biodiversity and ecology.
Keywords: YOLO (You Only Look Once)
Document type: info:eu-repo/semantics/masterThesis
Issue Date: 24-Jan-2025
Publication license: http://creativecommons.org/licenses/by-nc-nd/3.0/es/  
Appears in Collections:Bachelor thesis, research projects, etc.

Files in This Item:
File Description SizeFormat 
TFM_Carlos_Romero_Matarin.docxMemoria editable9,68 MBMicrosoft Word XMLView/Open
cromeromatTFM0125memoria.pdfMemoria del TFM2,44 MBAdobe PDFThumbnail
View/Open
cromeromatTFM0125codigo.pdfCódigo principal del TFM1,82 MBAdobe PDFThumbnail
View/Open
Share:
Export:
View statistics

This item is licensed under aCreative Commons License Creative Commons