Deep Learning-Based Semantic Segmentation Models for Crane Detection in Post-Earthquake UAV Imagery
Engineering and Technology Journal · 30.06.2026 · 11 (06)
DOI: 10.47191/etj/v11i06.55 ↗Kurz gefasst
Fachartikel in der Zeitschrift Engineering and Technology Journal, online erschienen am 30.06.2026. Verfasst von Salih Enes Şeker und 1 weiteren Autor:innen.
Abstract (Originalsprache)
The rapid identification of heavy machinery in post-disaster environments can significantly support recovery and reconstruction operations. This study presents a comparative evaluation of deep learning based semantic segmentation models for crane segmentation in UAV imagery acquired after the February 6, 2023 earthquakes in Kahramanmaraş, Türkiye. Five semantic segmentation architectures, namely SegFormer, DeepLabV3+, SegNet, FCN and U-Net, were trained and evaluated under identical experimental conditions. The results showed that DeepLabV3+ achieved the best overall quantitative performance with an IoU of 66.1%, mIoU of 81.0%, and Dice score of 78.9%. The findings demonstrate that modern segmentation architectures provide effective solutions for crane segmentation in complex post-earthquake UAV imagery and have significant potential for supporting disaster recovery and situational awareness applications.
Zitieren
APA
Şeker, S. E., & Çankaya, İ. (2026). Deep Learning-Based Semantic Segmentation Models for Crane Detection in Post-Earthquake UAV Imagery. Engineering and Technology Journal, 11(06). https://doi.org/10.47191/etj/v11i06.55
Vancouver
Şeker SE, Çankaya İ. Deep Learning-Based Semantic Segmentation Models for Crane Detection in Post-Earthquake UAV Imagery. Engineering and Technology Journal. 2026;11(06). doi:10.47191/etj/v11i06.55
BibTeX
@article{eker2026,
author = {Şeker, Salih Enes and Çankaya, İlyas},
title = {Deep Learning-Based Semantic Segmentation Models for Crane Detection in Post-Earthquake UAV Imagery},
journal = {Engineering and Technology Journal},
year = {2026},
volume = {11},
number = {06},
doi = {10.47191/etj/v11i06.55}
}
EUVIMED