qAicedrone-­Roads. A Robust Tool for Road Marking Extraction Using Aerial Photogrammetry and U-­Net

[EN] Efficient and accurate road marking detection is essential for infrastructure maintenance, traffic management, and the develop- ment of digital twins for autonomous mobility. However, most existing methods rely on orthomosaics or single-­image detections, which suffer from geometric distortions...

Descripción completa

Detalles Bibliográficos
Autores: Barbero-García, Inés|||0000-0003-1049-7586, Marqués-Mateu, Ángel|||0000-0003-1343-103X, Martínez-Lastras, S., Hernande-López, David
Tipo de recurso: artículo
Fecha de publicación:2025
País:España
Institución:Universitat Politècnica de València (UPV)
Repositorio:RiuNet. Repositorio Institucional de la Universitat Politécnica de Valéncia
Idioma:inglés
OAI Identifier:oai:riunet.upv.es:10251/227734
Acceso en línea:https://riunet.upv.es/handle/10251/227734
Access Level:acceso abierto
Palabra clave:Digital twin
Infrastructure monitoring
Multiview photogrammetry
Road marking Detection
Semantic segmentation
UAV photogrammetry
09.- Desarrollar infraestructuras resilientes, promover la industrialización inclusiva y sostenible, y fomentar la innovación
11.- Conseguir que las ciudades y los asentamientos humanos sean inclusivos, seguros, resilientes y sostenibles
Descripción
Sumario:[EN] Efficient and accurate road marking detection is essential for infrastructure maintenance, traffic management, and the develop- ment of digital twins for autonomous mobility. However, most existing methods rely on orthomosaics or single-­image detections, which suffer from geometric distortions and occlusions and have limited semantic insight. To address these limitations, this study introduces qAicedrone-­Roads, an open-­source tool integrated into QGIS that enables the automatic detection, classifi- cation, and mapping of road markings from UAV-­based photogrammetric imagery. The methodology combines U-­Net-­based semantic segmentation, a multiview photogrammetric approach, and alignment with a national road marking catalog to enhance geometric accuracy and assign semantic labels. Applied to a real-­world case study, the tool achieved high precision, with F1-­ scores of 0.92 for nonlinear and 0.93 for linear markings, outperforming traditional single-­v iew Computer Vision (CV) methods. These results demonstrate the tool's robustness and accuracy in complex urban environments, enabling the efficient generation of detailed road marking datasets. By facilitating the scalable and reproducible creation of digital twins, qAicedrone-­Roads sup- ports smart infrastructure monitoring and sustainable urban mobility planning.