Flood and Landslide
Early Detection System

Sri Lanka frequently experiences floods and landslides, which pose serious risks to public safety, infrastructure, and the economy. Flooding, driven by heavy rainfall, disrupts key sectors such as transportation and agriculture, while landslides in mountainous areas damage homes and block roads. The tragic 2016 Aranayake landslide, which claimed many lives, underscores the need for more reliable and effective disaster prediction systems.

Currently, disaster management in Sri Lanka relies on traditional methods such as static data analysis and Geographic Information Systems (GIS) to assess risk areas. While these approaches provide valuable insights into historical patterns, they often fall short in predicting the dynamic nature of disasters in real time. This results in slower response times and less precise risk forecasts, particularly in rural and remote regions.

To overcome these limitations, new technologies like machine learning and IoT sensor networks offer transformative potential. By utilizing real-time data from sensors monitoring rainfall, water levels, and soil stability, predictive models can continuously update risk assessments, allowing for more accurate and timely forecasts. This integration of advanced technology would enhance early warning systems, facilitate quicker evacuations, and significantly reduce the impact of floods and landslides on communities and infrastructure.

Proposed Problem

The frequent and severe natural disasters like floods and landslides in Sri Lanka pose significant threats to public safety and economic stability

The current disaster prediction and management systems in Sri Lanka are hindered by poor sensor accuracy and unreliable data transfer, especially in remote areas. This research aims to improve disaster response by integrating machine learning (ML) and Internet of Things (IoT) technologies with Long Range (LoRa) communication to create a scalable and reliable early detection system for floods and landslides.

The methodology of this research focuses on developing an advanced disaster response system using machine learning (ML) and Internet of Things (IoT) technologies to predict and detect floods and landslides in Sri Lanka. The system employs Long Range (LoRa) technology to create a Scalable Wireless Sensor Network that ensures reliable data transfer across rural areas. This network collects real-time environmental data, which is then processed by the flood prediction model to enhance its accuracy continuously. For landslide detection, the system uses image processing techniques to analyze satellite images for early signs such as cracks. The integrated alarm system provides precise, real-time alerts, significantly improving response times. The models are trained using datasets specific to Sri Lanka, ensuring high relevance and accuracy. Additionally, a web application delivers location-specific disaster notifications, advancing disaster management and preparedness. This approach not only aims to safeguard communities but also to protect economic assets, addressing the urgent need for improved disaster prediction and response systems in Sri Lanka. The methodology’s flexibility and scalability allow it to be applied in other regions with similar geographic and infrastructural challenges, promoting sustainable development and disaster resilience.