Flood Detection and Water Estimation in Aerial Imagery Using Deep Learning

Automated flood damage analysis is the focus of this project, utilizing deep learning, computer vision techniques, and satellite/aerial photos. The primary goal is to create a segmentation model that accurately identifies water patches in flood photos, surpassing traditional methods with advanced algorithms like Multi-Layer Perceptron (MLP) classification and the U-Net segmentation model. Emergency response teams stand to benefit greatly from this research. The precise identification of flooded areas will enable effective planning, resource allocation, damage evaluation, and post-disaster recovery operations. Special attention is given to estimating the flood-affected area, leveraging the developed MLP classification and U-Net segmentation methods, leading to improved flood management strategies and a reduction in the impact of flood-related disasters. Through the combination of deep learning and computer vision, this project aims to revolutionize flood damage assessment, providing essential tools for faster and more accurate decision-making. By mitigating the effects of floods, it paves the way for safer and more resilient communities.