Recommended Equipment
While modern smartphones offer convenience and rapid image acquisition, DSLR cameras are typically preferred due to their higher image quality, greater sensor sensitivity, and superior exposure control. These advantages are particularly important in underground environments, where lighting conditions are limited and reliable low-light performance is essential.
To minimize motion blur, especially during long exposure times, cameras are typically mounted on stable tripods. Many modern DSLR systems also feature live-view functionality and remote triggering, which improve usability in confined underground spaces by enabling more precise framing and focusing (Figure 1).
Image Acquisition
The imaging setup in underground environments differs from surface imaging due to limited space, reduced lighting conditions, and tight time constraints. Reliable photogrammetric reconstruction requires systematic image acquisition, considering camera distance, viewing geometry, and image overlap.
An image overlap of approximately 80% between consecutive images is required to enable robust feature matching and stable reconstruction. The distance to the rock surface does not need to remain constant but should be kept within a controlled range to ensure consistent overlap and resolution. Images should be captured from multiple viewpoints to achieve full geometric and textural coverage, including both perpendicular and oblique angles to effectively document complex shapes and discontinuities.
If the basic prerequisites are met, imaging on site can be flexibly adapted according to the project requirements (Figures 2 and 3). Depending on the complexity, size, and accessibility of the excavation, datasets may range from a few dozen to several hundred photographs.
Handling Lighting and Exposure
Proper illumination is essential for high-quality image acquisition in underground environments. Tunnel faces and rock walls are typically illuminated using portable floodlights or LED lighting systems. Flash photography is generally avoided, as it can produce inconsistent shadowing due to irregular rock surfaces, generate visual artifacts, and is usually not strong enough to illuminate the entire scene.
Instead, portable LED arrays are preferred. They provide continuous and stable illumination, operate efficiently on battery power, and are well-suited for field conditions. The use of multiple light sources helps reduce shadows and ensures more uniform lighting (Figure 4).
Registering the 3D model
Two primary approaches are commonly used in underground environments:
- Ground Control Points (GCPs) are surveyed reference markers with known coordinates. These points must be clearly visible in the images and are typically:
- Spray-painted targets on rock surfaces, or
- Pre-installed surveying markers designed for repeat use, often also used in deformation monitoring systems.
The coordinates of GCPs are usually measured using a total station or similar high-precision surveying equipment. Ideally, GCPs should be distributed throughout the entire area of interest to ensure a well-constrained model (Figure 5). However, in practice, access limitations often restrict their placement. A typical underground configuration includes at least three well-distributed points on both sides of the tunnel wall at accessible heights.
- Range Poles and Scaling Devices allow scaling and basic orientation without requiring survey team involvement or total station measurements during image acquisition (Figure 6). The range pole provides:
- A known scale reference for dimensional control
- Vertical orientation for consistent spatial reference
- A practical basis for model scaling when full georeferencing is not available
Directional orientation (north correction) can be incorporated using a compass measurement or derived from the known alignment of the tunnel or drift axis. Although this approach is less precise than full GCP-based georeferencing, it is highly efficient where rapid data acquisition is required.
Laser Scanner Data Integration
These approaches are complementary. In addition to photogrammetric data, external 3D datasets can be integrated into the workflow. Laser scanning (LiDAR) data is particularly valuable in underground environments due to its high accuracy and dense spatial coverage. Point clouds generated from LiDAR systems can be imported through the E57 lidar point cloud file format (Figure 7).
Training Requirements
The workflow is designed to be accessible without dedicated on-site training. Images can be captured with safety considerations and the requirements for 3D model generation in mind. The generation of 3D models is largely automated using the software MultiPhoto within ShapeMetriX. The system is capable of producing high-quality reconstructions with minimal manual intervention. The workflow typically runs automatically, while the user adjusts the model resolution depending on the desired level of accuracy. Scaling and georeferencing are integrated during processing. A user-guided interface supports the assignment of control points to corresponding image features. The final output is a real-world scaled 3D model of the excavation.
Geological Mapping
The generated 3D models are subsequently used for geological and geotechnical interpretation in the Analyst, a dedicated module within ShapeMetriX. Outcropping geological structures such as traces and surfaces are mapped directly on the 3D model (Figure 8). From these digital annotations, key geological parameters are derived, including:
- Dip direction and dip angle
- Spatial position of discontinuities
- Persistence and trace length
- Spacing and frequency of discontinuities
Mapping can be performed either interactively or supported by semi-automatic detection tools that assist in identifying geological eatures. Advanced interpolation and forward projection modelling tools are also available for planar and undulating discontinuities. Instant visualizations, including stereonet plots and joint spacing analyses, can be generated and customized according to specific requirements. Additional features such as scanline mapping, discontinuity clustering, and statistical evaluation support geotechnical interpretation.
Summary
Underground photogrammetric data acquisition enables the efficient generation of spatially accurate 3D rock mass models under challenging conditions. Through systematic imaging, controlled lighting, and appropriate scaling or georeferencing, highly reliable digital representations of geological structures can be produced using largely automated workflows.
These models enable detailed structural analysis, support geotechnical classification, and facilitate the early identification of unstable zones. They also provide a consistent framework for documenting excavation geometry and monitoring geological evolution over time.
Photogrammetry provides a fast, flexible, and precise method for documenting underground conditions, delivering reproducible, high-quality data while enhancing operational safety through reduced personnel exposure to hazardous areas.