EUVIMEDEUVIMED entdecken
Crossref

Semantic Segmentation of Flood-Affected Areas, Deep Learning Models (SegNet, UNet, FCN32) and random forest regression: A Remote Sensing Approach For Disaster Monitoring

Rajendra Kattunga

Journal of Software Quality Engineering and DevOps · 2026 · 1 (1) · 1-12

DOI: 10.55124/jsqd.v1i1.105 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Software Quality Engineering and DevOps, 2026. Verfasst von Rajendra Kattunga.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Floods, One of the most destructive natural catastrophes in the globe, it damages infrastructure and has major social effects, especially in disaster-prone areas such as Indonesia. This research addresses the vital requirement for precise and prompt flood detection and monitoring systems through advanced remote sensing and machine learning technologies. This study evaluates the performance of deep learning frameworks, especially for semantic segmentation of flood-affected areas, DeepLabv3 uses a variety of pre-trained backbone models, such as ResNet50, EfficientNet-B4, and MobileNet using the Flood Net dataset of 398 reference samples. In addition, a comparative analysis of state-of-the-art frameworks, including SegNet, UNet, and FCN32, demonstrates their effectiveness in automatic flood area detection, with SegNet achieving the best performance at 88% accuracy. The research methodology includes key parameters including rainfall intensity, flooded area size, water depth measurements, and segmentation accuracy parameters including F1-score, recall, accuracy, and mIoU. The incorporation of synthetic aperture radar (SAR) technology and satellite-based remote sensing provides extensive monitoring capabilities over large geographical areas. This multi-faceted approach, which combines integrated techniques such as unmanned aerial vehicles, deep learning architectures, and stochastic forest regression, provides a strong foundation for improving disaster recovery, flood forecasting systems, and risk reduction strategies in vulnerable communities. Keywords: Flood Detection, Semantic Segmentation, Deep Learning, Remote Sensing, DeepLabv3, UAV Surveillance, Disaster Management, SAR Images, Machine Learning, Computer Vision

Zitieren

APA

Kattunga, R. (2026). Semantic Segmentation of Flood-Affected Areas, Deep Learning Models (SegNet, UNet, FCN32) and random forest regression: A Remote Sensing Approach For Disaster Monitoring. Journal of Software Quality Engineering and DevOps, 1(1), 1-12. https://doi.org/10.55124/jsqd.v1i1.105

Vancouver

Kattunga R. Semantic Segmentation of Flood-Affected Areas, Deep Learning Models (SegNet, UNet, FCN32) and random forest regression: A Remote Sensing Approach For Disaster Monitoring. Journal of Software Quality Engineering and DevOps. 2026;1(1):1-12. doi:10.55124/jsqd.v1i1.105

BibTeX
@article{Kattunga2026,
  author = {Kattunga, Rajendra},
  title = {Semantic Segmentation of Flood-Affected Areas, Deep Learning Models (SegNet, UNet, FCN32) and random forest regression: A Remote Sensing Approach For Disaster Monitoring},
  journal = {Journal of Software Quality Engineering and DevOps},
  year = {2026},
  volume = {1},
  number = {1},
  pages = {1--12},
  doi = {10.55124/jsqd.v1i1.105}
}
Kostenlos starten →