Water Level Detection and Flood Early Warning System Using Image Processing

Abstract

Image processing is a crucial method in modern technology, enabling computers to analyze and extract information from images or videos. This study focuses on the application of image processing technology to detect river water levels using CCTV cameras as part of a flood early warning system in a smart city. The YOLO (You Only Look Once) algorithm is utilized for real-time object detection, such as water levels, aiming to enhance prediction accuracy. This implementation is expected to provide richer visual data compared to traditional sensors. The study involves designing and testing a system that integrates hardware (CCTV cameras and high-spec computers) and software such as OpenCV and Python. Data in the form of river images is processed using image processing algorithms to analyze water levels in real-time. The system's performance is evaluated in terms of accuracy, precision, recall, and processing speed (FPS), as well as the environmental impact on detection results. The results indicate that the YOLO-based image processing system achieves high accuracy in detecting water levels. Additionally, the system is capable of sending early warnings via digital notifications, allowing more time for disaster mitigation. These findings suggest that image processing-based systems offer practical, efficient, and cost-effective solutions to support smart city technologies.