Executive Summary : | Landslides are devastating and cause significant loss of life, property, and infrastructure. Meghalaya, India's rainiest state, experiences frequent landslides during the monsoon season. To address this issue, a project is proposed to use Machine Learning (ML) techniques to generate a Landslide susceptibility Map (LsM) for Meghalaya, a network of Internet of Things (IoT) to monitor active landslides, and implement an Early Warning system (EWs) to alert people in the risk area and authorities about potential landslides. The ML model will be trained using remote sensing data from sources like U.s. Geological Earth Explorer, GsI Bhukosh, NAsA-UsDA Enhanced sMAP Global soil Moisture Data, Global Land Cover Characterization (GLCC), and UsGs. The model will be validated by accuracy, Kappa, and the area under the receiver operating characteristic curve (AUC-ROC). The proposed IoT architecture for landslide monitoring and EWs includes four layers: edge/things layer, edge computing layer, cloud computing and storage layer, and application and monitoring layer. The edge computing layer collects data from sensors, transmits it wirelessly to the edge computing layer, processes it for real-time warnings, and passes it to the cloud computing and storage layer. The computing and storage layer analyzes the collected data, while the application and monitoring layer provides a user-friendly interface for visualization and alerting. This project aims to mitigate landslide loss in Meghalaya, contribute to existing knowledge, and be applicable to other regions. |