Monitoring Method for Road Waterlogging Based on Improved YOLOv8

Abstract

Flooding disasters occur frequently in China, especially on roads, which seriously affect people's normal travel and even threaten their lives. The current technologies for monitoring road waterlogging are inefficient, and there is an urgent need for an efficient method to monitor road waterlogging. Accurate monitoring of road waterlogging is helpful for the government to issue policies and personnel to take preventive measures. Therefore, this article proposed a real-time monitoring method for road waterlogging based on improved YOLOv8. Through the YOLOv8 algorithm, a convolutional block attention module (CBAM) attention mechanism was added to the neck structure network to enhance the important features of waterlogging areas, suppress general features, and improve the accuracy of identifying road waterlogging. In addition, perspective transformation and pixels were used to calculate the waterlogging area. The article studied the road waterlogging in the new campus of Hebei University of Engineering. The results show that the accuracy of this method reaches 93.83%, which can accurately identify the road waterlogging surface and output the waterlogging area in real time, meeting the monitoring needs.

Keywords

• waterlogging disaster;deep learning;attention mechanism;road waterlogging