AI-Powered Drainage Monitoring System Using Computer Vision and IoT Sensors for Proactive Flood Prevention

Abstract

Flooding is one of the most significant results of urban drainage blockages; the impact is often property damage, health problems, or economic losses. Traditional manual inspections of urban drainage systems require several hours of labor at each location and frequently do not provide monitoring of obstructions in an efficient manner. The purpose of this study was to develop a fully AI-enabled drainage monitoring system to monitor urban drainage systems in real-time and proactively prevent urban flooding. The AI-Enabled Drainage Monitoring System was developed using a Raspberry Pi as the main processor, a USB webcam to acquire images of drainage conditions, an ultrasonic sensor for continuous monitoring of drainage systems, and a water level sensor to measure actual water levels in the drainage system. An AI-based image classification model (ICM) was created to classify drainage conditions as either clear, partially blocked, or fully blocked. A web-based dashboard using Flask provides real-time monitoring data, alert notifications, historical records, and weather forecasts; this dashboard allows Local Government Units (LGUs) to make more informed decisions and also provides selected historical data to the public to help increase public awareness of the importance of monitoring urban drainage systems. The study employed a developmental research design and the Agile Software Development Life Cycle (SDLC). Results indicate that the system can effectively detect drainage blockages and generate timely alerts, demonstrating a scalable and cost-effective solution for urban flood risk mitigation.