Real-Time Monitoring and Safety Warning Technology of Construction Site in Engineering Management BIM Combined with YOLOv8

With the continuous expansion of construction projects and the increasing complexity of construction sites, safety management during construction is facing unprecedented challenges. Traditional safety management methods often rely on manual inspections, which are inefficient and prone to ignoring potential risks, making it difficult to achieve real-time, dynamic monitoring and early warning. To solve this problem, this study introduces a construction site real-time monitoring and safety warning technology for engineering management BIM (Building Information Modeling) combined with YOLOv8 (You Only Look Once Version 8). First, a three-dimensional digital model of the construction site is established based on BIM (Building Information Modeling) technology, and the on-site video is analyzed in real-time through the YOLOv8 deep learning model to automatically detect abnormal conditions of construction personnel, equipment and dangerous factors. Next, this paper combines the spatial information in the BIM model with the detection results of YOLOv8 to conduct a multi-dimensional safety risk assessment. Finally, this paper uses the preset safety thresholds and rules to generate safety warning information and push it to relevant personnel in real-time so that they can take timely countermeasures. Experimental results show that YOLOv8 performs well in various indicators: in terms of target detection accuracy, YOLOv8’s IoU (Intersection over Union) and mAP (Mean Average Precision) are 0.85 and 0.92 respectively, which are significantly higher than YOLOv4 and Faster R-CNN (Faster Region Convolutional Neural Network). In terms of safety warning response time, the average response time of YOLOv8 is only 0.32 seconds. In terms of safety hazard recognition rate, the average recognition rate of YOLOv8 is 91.81%. In the above data conclusions, YOLOv8 performs well in construction site target detection and safety monitoring tasks, with high accuracy, real-time and reliability, and is suitable for safety monitoring systems in complex construction environments.