Інтегрована система прогнозування паводків і картування затоплення на основі супутникових даних та методів машинного навчання

Abstract

The intensive floods can be very dangerous for our infrastructure and our lives. This study addresses the growing risk of floods caused by climate change and the limitations of traditional hydrological methods, which often fail to capture spatial variability in precipitation and environmental changes. The proposed approach integrates high-resolution satellite data with machine learning techniques to improve early flood prediction. By integrating forecasting and mapping modules, the system creates a feedback loop where predictions guide satellite observations and observed flood maps help refine the models. This approach improves the accuracy of early warning systems and enhances long-term adaptation to flood risks