Three Stage Approach for Waterlogging in Cable Tunnel Using Machine Learning and IOT

Urban underpasses and cable tunnels in Indian cities frequently experience rapid waterlogging during intense rainfall, causing traffic disruption, vehicle damage, and serious safety risks to pedestrians and motorists. This paper presents a three‑stage smart underpass waterlogging management system that combines embedded sensing, Internet of Things connectivity, and basic machine‑learning‑based computer vision for human safety. Ultrasonic water‑level, rain, and contact sensors interface with an Arduino‑based controller, which classifies the situation into safe, caution, and danger zones, drives a 16×2 LCD for vehicle‑specific guidance, and transmits real‑time data to a cloud platform through a NodeMCU module. When the level exceeds a critical threshold, a relay‑controlled pump is automatically activated for drainage, while a camera feed is processed using Python, OpenCV, and a Haar Cascade classifier to detect humans in the flooded region and trigger emergency alerts. Prototype results demonstrate reliable level classification, timely activation of the pump, and successful human detection in controlled flood scenarios, indicating that the proposed system offers a low‑cost, scalable approach to improving safety and resilience of urban underpass infrastructure. ​