Scholarlink

<p>The integration of flood detection and warning systems utilizing machine learning and the Internet of Things (IoT) represents a significant advancement in flood risk management. Floods, being unpredictable natural disasters, necessitate timely and effective detection systems to mitigate potential damages. Traditional flood control measures, while essential, are often costly and cannot fully eliminate flood risks. Therefore, a non-structural approach that includes real-time monitoring and alerts is critical. This article proposes a comprehensive flood detection, warning, and prediction system that employs IoT devices to monitor water levels in various water bodies, including rivers, dams, canals, and lakes. The IoT nodes continuously collect data on water levels and environmental conditions. When the water level surpasses a predetermined threshold, alerts are dispatched to users within a 5 km radius. An Android application displays vital information such as temperature, water levels, flood locations, and predictions. The system’s effectiveness is underscored by the performance of the Linear Regression algorithm, which achieves an accuracy rate of 79.01%. This is notably higher than other algorithms like Decision Tree and Naïve Bayes in the context of flood detection using IoT. Additionally, the system boasts a recall rate of 91.00%, indicating its strong capability to correctly identify flooding events, while maintaining a precision rate of 84.30%.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A flood expert system using machine learning and IoT: warning, detection, and prediction

  • 期刊论文

  • 发表日期：

    2025年1月20日

查看全文
  • Soleyman Nezhadbasaidu,
  • Mehdi Gheisari,
  • Alireza Kheyrkhah,
  • Mohammad Hossein Modirrousta,
  • Xiuqing Wang,
  • Sherif Moussa,
  • Hemn Barzan Abdalla,
  • Belal Abuhaija

摘要

The integration of flood detection and warning systems utilizing machine learning and the Internet of Things (IoT) represents a significant advancement in flood risk management. Floods, being unpredictable natural disasters, necessitate timely and effective detection systems to mitigate potential damages. Traditional flood control measures, while essential, are often costly and cannot fully eliminate flood risks. Therefore, a non-structural approach that includes real-time monitoring and alerts is critical. This article proposes a comprehensive flood detection, warning, and prediction system that employs IoT devices to monitor water levels in various water bodies, including rivers, dams, canals, and lakes. The IoT nodes continuously collect data on water levels and environmental conditions. When the water level surpasses a predetermined threshold, alerts are dispatched to users within a 5 km radius. An Android application displays vital information such as temperature, water levels, flood locations, and predictions. The system’s effectiveness is underscored by the performance of the Linear Regression algorithm, which achieves an accuracy rate of 79.01%. This is notably higher than other algorithms like Decision Tree and Naïve Bayes in the context of flood detection using IoT. Additionally, the system boasts a recall rate of 91.00%, indicating its strong capability to correctly identify flooding events, while maintaining a precision rate of 84.30%.