Satellite Image-Based Tsunami detection with Automated Alert and Evacuation System

This system proposes a real-time tsunami prediction and alert system that integrates live satellite imagery with machine learning concepts to analyze the Indian ocean conditions and detect tsunami events accurately. The system utilizes past augmented images of the Indian Ocean from the NASA Earth data, and uses them to train the model using a ResNet architecture and also with architectures like MobileNet anf EfficientNet. It then compares the past data with the live data from satellite using CNN concepts to identify changes in water conditions and other critical indicators of potential tsunamis. Upon detection of a potential tsunami, the system creates a geofence to identify the people around a 500m radius and automatically disseminates alerts through Email notifications along with the location of the nearest safe center, ensuring immediate communication without the need for specific app installations. Additionally, it provides specific evacuation routes to guide the affected population to safety promptly. The project also includes a detailed study of existing tsunami detection and alert systems, highlighting the limitations and areas for improvement.