Título:	Mapping stains on flat roofs using semantic segmentation based on deep learning
Autor(es):	Santos, Lara Monalisa Alves dos Lescano, Leonardo Rabero Higa, Gabriel Toshio Hirokawa Zanoni, Vanda Alice Garcia Silva, Lenildo Santos da Alvarez Mendoza, Cesar Ivan Pistori, Hemerson
ORCID:	https://orcid.org/0000-0002-8022-2513 https://orcid.org/0009-0004-3125-9696 https://orcid.org/0009-0006-6771-0076 https://orcid.org/0000-0003-2629-4214 https://orcid.org/0000-0001-5099-6123 https://orcid.org/0000-0001-5629-0893 https://orcid.org/0000-0001-8181-760X
Afiliação do autor:	University of Brasilia Dom Bosco Catholic University Dom Bosco Catholic University University of Brasilia University of Brasilia University of Augsburg, Centre for Climate Resilience Salesian Polytechnic University, Environmental Research Group for Sustainable Development (GIADES) Dom Bosco Catholic University Federal University of Mato Grosso do Sul
Assunto:	Aprendizagem profunda Inspeção predial Drones Visão computacional
Data de publicação:	18-Dez-2024
Editora:	Elsevier Ltd.
Referência:	SANTOS, Lara Monalisa Alves dos et al. Mapping stains on flat roofs using semantic segmentation based on deep learning. Case Studies in Construction Materials, [S. l.], v. 22, e04106, 2024. DOI: https://doi.org/10.1016/j.cscm.2024.e04106. Disponível em: https://www.sciencedirect.com/science/article/pii/S2214509524012580?via%3Dihub. Acesso em: 11 mar. 2025.
Abstract:	Moisture stains indicate ongoing degradation processes and may reveal areas of the roof slab where water infiltration occurs, compromising the performance and durability of the building system. During inspections of roofing systems, an inspector’s field of vision differs from that of drones during overflights. As a result, traditional inspections might not always detect the presence and severity of stains, making maintenance on flat roofs a complex task. In this context, this experimental study aims to analyze deep learning-based semantic segmentation with images obtained from drones to map and monitor damp patches during automated building inspections of flat roof systems. The research tested two convolutional neural networks for semantic segmentation: the Fully Convolutional Network (FCN) with a ResNet50 backbone and DeepLabV3 with a ResNet101 backbone, as well as a transformer-based deep artificial neural network called SegFormer with a MiT-B1 backbone. We evaluated three optimizers for each model—Adam, Adagrad, and SGD—along with learning rates of 1e-2, 1e-3, and 1e-4. The models were compared using four performance metrics. The FCN, optimized with Adagrad at a learning rate of 1e-2, showed the best results. The average metrics obtained in this case were as follows: precision: 79.69 %, recall: 67.81 %, F-score: 73.09 %, and Intersection over Union (IoU): 57.70 %.
Unidade Acadêmica:	Faculdade de Arquitetura e Urbanismo (FAU) Departamento de Tecnologia em Arquitetura e Urbanismo (FAU TEC) Faculdade de Tecnologia (FT) Departamento de Engenharia Civil e Ambiental (FT ENC)
Programa de pós-graduação:	Programa de Pós-Graduação em Arquitetura e Urbanismo
Licença:	This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
DOI:	https://doi.org/10.1016/j.cscm.2024.e04106
Aparece nas coleções:	Artigos publicados em periódicos e afins

Os itens no repositório estão protegidos por copyright, com todos os direitos reservados, salvo quando é indicado o contrário.