Skip to content
SDGs Center UDINUS

Evidence record · UDINUS · SDGs

UDINUS Researchers Develop Satellite-Based Flood Mapping Technology to Support Disaster Resilience and the SDGs

Lembaga Penelitian dan Pengabdian Masyarakat 2026 Published 2026-11-16

Evidence record · UDINUS · SDGs

Unit
Lembaga Penelitian dan Pengabdian Masyarakat
Period
2026
Goals
Record
a1a9c44c-cb5e-4021-8e92-b3abe990d431
Published
2026-11-16

Researchers from Universitas Dian Nuswantoro (UDINUS) have developed an innovative satellite-based flood mapping approach integrating remote sensing and machine learning technologies. The research aims to improve flood inundation detection in complex wetland environments, supporting disaster risk management and the achievement of the United Nations Sustainable Development Goals (SDGs).

The study, entitled "Hybrid SAR-Optical Remote Sensing for Flood Inundation Mapping: Feature Contribution Analysis in a Wetland Environment," was published in Statistics, Optimization & Information Computing, Volume 16, Number 1 (2026), pages 619–640, by International Academic Press.

The research was conducted by Desy Ika Puspitasari, Edi Noersasongko, Purwanto, Moch Arief Soeleman, and Catur Supriyanto, with affiliations to the Faculty of Computer Science, Universitas Dian Nuswantoro.

Flooding remains a significant environmental challenge in Indonesia, particularly in low-lying regions characterized by extensive wetlands, high rainfall, and river sedimentation.

The researchers focused on Banjar Regency, South Kalimantan, an area frequently affected by seasonal flooding. Accurate flood inundation mapping is essential for identifying affected areas, supporting disaster response, and strengthening regional disaster preparedness.

To address the limitations of conventional satellite imaging, the research introduced a hybrid Synthetic Aperture Radar (SAR) and optical remote sensing approach, combining Sentinel-1 radar imagery with Sentinel-2 satellite data.

The proposed method integrates advanced water detection indices, including the Automated Water Extraction Index (AWEI), with machine learning algorithms to improve water classification and flood mapping reliability.

Four machine learning models were evaluated: Random Forest, Support Vector Machine (SVM), Logistic Regression, and XGBoost.

The results showed that the SVM model achieved strong water classification performance, with 98% overall accuracy, a Kappa coefficient of 0.97, and an F1-score of 0.98.

The study also demonstrated that integrating radar and optical data improves classification stability in complex wetland environments, where cloud cover and vegetation can limit conventional satellite observations.

These findings provide a valuable methodological foundation for more reliable flood inundation analysis and disaster-related spatial monitoring.

This publication highlights UDINUS's contribution to developing innovative technologies for environmental monitoring and disaster risk reduction.By combining satellite remote sensing with machine learning, the researchers demonstrate how advanced computational methods can improve the reliability of flood mapping in challenging geographical environments.

The findings provide opportunities for further research and the development of data-driven tools supporting disaster preparedness, environmental management, and sustainable regional development.

Through this scientific contribution, Universitas Dian Nuswantoro continues to advance technological innovation and research supporting disaster resilience and the United Nations Sustainable Development Goals.

Reference https://iapress.org/index.php/soic/article/view/4067/1937