Flood Relief Land Segmentation Path Mapping Tool Using U-Net Architecture

Land cover data provides insights into the distribution of various land and water types within a region, including forests, wetlands, impermeable surfaces, farmland, and more. These diverse land cover types can be managed in various ways based on their characteristics. In this study, we explore the use of satellite and drone imagery to determine land cover and manage disasters. The management of natural or man-made disasters involves planning, response, and recovery to minimize their impact on people, infrastructure, and the environment. This paper discusses a web application utilizing CNN U-Net architecture and shortest path algorithms for land cover segmentation and classification, with a focus on flood relief. We demonstrate the ability to identify flooded regions and optimize rescue routes with an accuracy of 82%.