HydroLeak

Demo: https://docs.google.com/file/d/1U7RqrJMRnAVW8p23aVt4xdNy-aXkYglN/preview

This project aims to develop an intelligent monitoring system for water pipeline infrastructure by leveraging Synthetic Aperture Radar (SAR) imagery and advanced deep learning techniques to detect leakages accurately and efficiently.

Water loss due to undetected leaks is a major global issue, leading to significant economic and environmental impacts. Traditional monitoring methods—such as manual inspections and in-ground sensors—are often costly, time-consuming, and limited in spatial coverage. To address these limitations, this project utilizes satellite-based SAR data, which has the advantage of operating in all weather conditions and during both day and night.

The system processes multi-temporal SAR imagery to identify subtle surface changes associated with underground water leaks, such as soil moisture variations and ground deformation. By applying deep learning models—particularly convolutional neural networks (CNNs) and temporal architectures—the project aims to automatically learn patterns indicative of leakage events.