high-grade, and waste domains are spatially registered with the physical bench surface, enabling intuitive real-time perception of grade distribution for on-site operational awareness (Figure 2). 2.3. Positioning Framework and Coordinate Transformation Accurate spatial correspondence between the digital ore-grade model and the physical bench environment is fundamental for reliable augmented-reality grade visualization and operational grade control. The positioning framework of the proposed system is established using Real-Time Kinematic Global Navigation Satellite System (RTK GNSS) measurements, in which satellite carrier-phase observations are combined with differential correction data transmitted from a fixed reference station to provide high-precision real-time positioning. GNSS observations are initially obtained in geodetic form, expressed as latitude, longitude, and ellipsoidal height relative to the reference datum. For compatibility with mine design and geological modeling workflows, these geodetic coordinates are transformed within the mobile integration hub into a projected Cartesian coordinate system consistent with the mine spatial reference framework. All positional and geological datasets in this study are referenced to the MGA GDA94 Zone 50 coordinate system, while vertical elevations are converted from ellipsoidal height to orthometric height using the AUSGeoid09 geoid model. This transformation establishes a common spatial reference between real-time GNSS positioning and the three-dimensional oregrade geometry generated in geological modeling software. Following projection, spatial calibration is required to relate the mine coordinate frame to the rendering coordinate system used by the augmented-reality environment. A rigid transformation consisting of translation and rotation parameters is applied to align the projected GNSS position with the Unity world coordinate frame in which the ore-grade model is visualized. The calibration parameters are determined using reference alignment constraints between the physical environment and the virtual model space, enabling consistent spatial registration during operation. The transformed three-dimensional position (X, Y, Z) is then continuously linked to the virtual camera pose within the AR visualization engine. This process ensures that the operator’s real-world movement at bench level is synchronously represented within the digital ore-grade model, preserving geometric consistency between physical and virtual environments. By maintaining a unified coordinate reference across GNSS positioning, geological modeling, and AR rendering, the framework supports stable spatial alignment required for bench-scale visualization and grade-control interpretation. 2.4. 3D Ore-Grade Model Preparation and Representation The 3D ore-grade model used in this study was developed using Maptek Vulcan software based on existing geological and grade control datasets. Rather than employing dense voxel-based block models, the system utilizes polygon-based grade representations that reflect operational grade control boundaries commonly used at the bench level. Each polygon represents a discrete ore classification zone, including low-grade, medium-grade, high-grade, and waste material. This representation reduces computational complexity and rendering load while maintaining sufficient spatial resolution for real-time field visualization. Polygon-based models are particularly well suited to AR applications, as they enable faster rendering and clearer visual interpretation on mobile hardware. All polygon models were generated in the same projected coordinate system as the GNSS data to ensure spatial consistency. The models were exported from Maptek Vulcan and changed to the Wavefront (OBJ) format in CloudCompare, preserving geometric accuracy and spatial positioning. Prior to integration into the AR environment, polygon meshes were optimized
RkJQdWJsaXNoZXIy MTM0Mzk2