Proof-of-concept pipeline for digital rock mass analysis using an inexpensive 360° camera

Source abstract

Rock mass characterisation, specifically characterising fracture distribution, density, and orientation, is crucial for a better understanding of stability and hydraulic properties. There are several methods for characterising fracture networks, such as manual mapping and remote sensing. This paper presents a proof-of-concept workflow for digital rock-mass analysis using a 360° imaging system photogrammetry combined with a three-dimensional Gaussian Splatting (3DGS). The method was tested on a tunnel drift (Site 1) and a 20-meter wall section (Site 2) in the Underground Research Laboratory of Aalto University (URLA). A compact 360° camera was used to acquire continuous panoramic image sequences of two sites in the tunnel. These images were processed into a 3D representation via photogrammetry combined with 3DGS, providing both accurate surface geometry and highly efficient and hyper-realistic real-time rendering. Fractures and discontinuity sets were extracted from the reconstructed surfaces using established methods such as 3D point cloud processing and semi-automatic clustering of fracture planes, allowing 3D fracture orientation measurements. The resulting orientations of discontinuity sets measured digitally from the 360° camera-based 3D point cloud differ 3.4° on average compared to previous measurements obtained from laser scanning and photogrammetry data. The 3D point cloud coloured by discontinuity sets was imported into Blender and visualised on the 3DGS. This enabled users to inspect the tunnel digitally, visualise structural features, and review spatial data in high resolution. The study demonstrates that a functional end-to-end digital rock-mass sensing pipeline can be assembled using an inexpensive 360° camera and existing software tools, offering a practical foundation for advancing automated rock-mass intelligence in future mining and tunnelling applications.