Optimized Flood Detection: Hyperparameter-Tuned CNN-SVM Fusion on SAR Imagery

Flood detection is critical for mitigating the impact of natural disasters, and Synthetic Aperture Radar (SAR) imagery has proven to be an effective tool due to its ability to penetrate clouds and provide detailed surface information. This study presents an optimized approach to flood detection employing Convolutional Neural Networks (CNN) and Support Vector Machines (SVM). The CNN extracts high-level features from SAR images, while the SVM classifier improves detection accuracy by learning from these features. A hyperparameter tuning strategy, including grid search and cross-validation, is applied to both the CNN and SVM components to achieve optimal performance. Experimental results on the SEN12-FLOOD dataset demonstrate that the CNN-SVM fusion model achieves an accuracy of 91%, precision of 89%, and recall of 88%, significantly outperforming traditional machine learning (ML) models and standalone CNN or SVM classifiers. This optimized fusion approach offers a robust solution for timely and accurate flood detection, enabling better decision-making in disaster response and management.