Skripsi
PEMANFAATAN SIDIK KOMPONEN UTAMA DALAM MENILAI GENANGAN BANJIR BERBASIS CITRA SYNTHETIC APERTURE RADAR
Flooding is a hydrometeorological disaster that predominantly occurs in Indonesia, resulting in significant material losses and casualties. Monitoring and mapping affected areas are crucial for flood risk management. This study investigates the utilization of Synthetic Aperture Radar (SAR) imagery from Sentinel-1, combined with change detection techniques and Principal Component Analysis (PCA) for flood inundation modeling. A log ratio-based change detection method is employed to identify temporal differences between images captured before and during flooding, while PCA is utilized to reduce speckle noise and enhance contrast between water bodies and land. The characteristics of intertemporal SAR images with different polarizations are evaluated, and the performance of PCA in modeling flood inundation is analyzed. The research findings indicate that the combination of LR1 and LR2 images is the most reliable in capturing flood changes, with a lower potential for interpretative errors compared to LR3 and LR4. The PCA output, when classified using a Random Forest Classifier (RFC), can predict flood inundation with an accuracy of 82.73%. These findings suggest that the integration of change detection techniques and PCA through SAR imagery is an effective approach for flood mapping, which can provide valuable insights for local governments in disaster mitigation.
No other version available