Application of construction site management based on digital twin and point cloud semantic segmentation technology

Area of Science:

• Construction Management
• Geospatial Technology
• Disaster Risk Reduction

Background:

• Construction sites face significant risks from extreme weather due to unsafe conditions.
• Current methods for assessing site risks (manual inspection, photography) are inefficient and time-consuming.
• There is a need for advanced technologies to improve construction site safety management.

Purpose of the Study:

• To integrate digital twin technology with point cloud semantic segmentation for construction site safety.
• To develop a system for systematic disaster preparedness against extreme weather.
• To optimize construction site layout and risk assessment using 3D point cloud data.

Main Methods:

• Utilized three-dimensional (3D) point cloud technology for construction site management.
• Developed a construction site point cloud identification system for material and machinery location.
• Applied RandLA-Net-based point cloud semantic segmentation within a digital twin framework.

Main Results:

• The developed system enables systematic disaster preparedness for construction sites.
• Location information of materials and machinery is acquired through the point cloud identification system.
• The digital twin model facilitates identification, location, and assessment of potential risk factors.

Conclusions:

• The integration of digital twins and point cloud segmentation significantly enhances construction site safety management.
• This approach optimizes site layout and risk mitigation strategies for extreme weather events.
• Practitioners can rapidly assess and address on-site risks, improving disaster prevention.