Aluizio Brito Maia

Detection of Flooded Areas Under Cloud Cover Conditions: Integration with Digital Elevation Models

Floods are hydrological processes directly associated with precipitation characteristics (intensity, frequency, and spatial distribution) and the geomorphometric features of the river basins where they occur. The detection of flooded areas using orbital images through remote sensing techniques requires robust models aimed at accurately identifying surface water. However, certain challenges hinder their detection. In optical sensors, the issue of cloud cover stands out, which leads to omission and inclusion errors of shadows during the image classification process. Several studies have analyzed the potential of image processing techniques for rapid disaster response, particularly for detecting flooded areas. However, few of these methods explore ways to extract information from cloud-covered images by integrating data of different natures. In this context, this dissertation proposes to develop a method for detecting flooded areas in images partially covered by clouds.