Real-time AR-based information retrieval for dynamic objects on construction sites

If real-time information on dynamic entities such as workers and equipment could be readily accessed, it would provide substantial benefits for construction site management. With the growing digitalization of the construction sector, the integration of Augmented Reality (AR) can further facilitate immediate access to such information. This study presents an enhanced AR framework that integrates computer vision–based detection and tracking with location and magnetometer data to support real-time monitoring of workers and equipment. At the core of this framework lies a matching process that associates image objects captured by the AR device with corresponding GPS-based location data, enabling retrieval of object-specific information directly within the camera view. The study incorporates directional consistency and entity grouping to improve matching stability and accuracy. Experimental evaluations demonstrated that the proposed approach achieved sufficiently high precision and recall to ensure stable information retrieval, while also operating reliably under real-time conditions.