Researchers from Incheon National University have developed an AI-based excavator tracking system designed to improve equipment monitoring on active construction sites.
The study, published in Automation in Construction, combines deep learning and a multi-camera strategy to improve tracking accuracy when excavators are blocked from view by other equipment or jobsite activity.
The system evaluates camera reliability in real time and automatically selects the clearest camera angle when visibility problems occur. Researchers also identified specific visibility thresholds where tracking accuracy begins to decline significantly.
According to the study, the technology could improve equipment operation tracking for productivity analysis, safety monitoring and carbon emissions reporting while reducing the need for additional cameras and manual oversight.
The researchers said the approach is designed to work with existing CCTV infrastructure commonly used on construction projects.
