'AI Flood Prediction System' Plays a Key Role in Preventing Heavy Rainfall Accidents

Gwangju City and AI Project Group 'Development and Support of Urban Problem Solutions'

Real-time Monitoring at Four Sites Including Wolchul Underpass

AI Algorithm Water Level Measurement Error Within 5mm

To prepare for heavy rains during the monsoon season, the ‘Artificial Intelligence (AI)-based Flood Prediction System’ established in underground roadways in the Gwangju area is playing a significant role in preventing flood accidents.

Gwangju Metropolitan City and the AI Industry Convergence Project Group announced on the 18th that they are demonstrating the ‘AI-based Flood Prediction and Monitoring System.’

This system was developed through the ‘2023 Urban Problem-Solving Solution Development and Support Project’ in collaboration between NH Networks Co., Ltd. and the AI startup Windy.

It is a complex local sensor system combining CCTV (closed-circuit television) and non-contact water level sensors. It was installed last October at Wolchul Underground Road in Yongjeon-dong, Buk-gu, Gwangju, and four surrounding areas to monitor water levels and flood situations in real time.

The information collected through monitoring is processed by an AI object recognition algorithm to identify objects such as people and vehicles, securing accurate water level data and reducing misrecognition.

This allows the water level measurement error range to be maintained below 5mm, enabling precise flood situation prediction, and is expected to play a major role in solving the recurring underground roadway flooding problems during the monsoon season.

Data collected from Wolchul Underground Road is visualized in real time through a geographic information system (GIS)-based control system and provided to citizens via a website. Additionally, meteorological information such as rainfall and precipitation probability from the Korea Meteorological Administration is linked to enhance the accuracy of flood predictions.

To support rapid and accurate responses to flood situations, electronic display boards have been installed at points 150m and 300m before entering the underground roadway to provide drivers with real-time flood information. This real-time warning system is expected to ensure citizens’ safety and help prevent secondary accidents.

The collected data will be continuously analyzed in connection with the AI data center within the Gwangju Advanced 3 District AI Cluster to further advance the Gwangju-type underground roadway flood prediction algorithm.

Kwack Jae-do, Acting Director of the AI Project Group, stated, “The AI-based flood prediction system will help prevent disasters through real-time data monitoring and accurate predictions, allowing citizens to live safely. We will do our best to develop and demonstrate AI technologies to solve various urban problems such as heavy rains.”

This content was produced with the assistance of AI translation services.