FREE TRIAL

Dense Reconstruction Presets and Their Influence on 3D Model Quality: Open-Pit Slope Example 

Geologists and engineers face a recurring question when working with photo-derived 3D models: how much detail is enough? Too coarse, and structural fabrics blur into the background. Too dense, and computation slows under the weight of unnecessary detail. 

For example, with just 33 aerial images captured under standard overlap and quality guidelines, ShapeMetriX reconstructs a mine slope that can be viewed, measured, and mapped in three dimensions. The real insight comes when the same slope is processed at different reconstruction settings, showing how model density transforms interpretation, texture fidelity, and analytical power in geotechnical workflows. 

ShapeMetriX offers flexible dense reconstruction presets, giving users full control over resolution during the dense reconstruction phase. This lets teams balance model fidelity, processing time, and performance—tailoring each dataset to the specific requirements of structural mapping or rock mass characterization. 

In this case study, we processed a bench in an open-pit mine under multiple presets, then compared the resulting 3D geometry, surface meshes, and textures side by side. The results show one simple truth: in geotechnical modeling, resolution defines interpretation.

Step-by-Step High-Resolution 3D Surface Reconstruction from Aerial Images

This case study demonstrates the full process, from aerial images to high-quality 3D model, using a single-bench open-pit slope as the subject. The objective: a model tailored for geological structural mapping.

Step 1: Aerial Images 

This example features 33 overlapping aerial images, taken from 4 rows in accordance with the 80% overlap and image quality suggestions, also presented in ShapeMetriX Field Procedures user guide. As shown below, load the images into MultiPhoto for the initial step of 3D model generation pipeline – Coarse Reconstruction.

Figure 1. Loaded images in configure project screen before coarse reconstruction.

Step 2: Coarse Reconstruction

This step establishes the relative orientations and positions of the cameras, mapping the general structure of the scene to a coarse 3D point cloud.  

Examine the coarse 3D model, set region of interest if needed and proceed to dense reconstruction. 

Figure 2. Preview of a 3D coarse reconstructed model: 3D point cloud and cameras (top), and Texture draped onto the point cloud (bottom). 

Step 3: Dense Reconstruction 

This step calculates the detailed object geometry including a high density point cloud and generates a detailed surface mesh with realistic textures. Dense reconstruction is essential for capturing terrain features, rock textures, and structural attributes needed for quantitative analysis and decision-making. 

Figure 3. Dense reconstruction 3D viewer with coarse 3D point cloud.
3.1 Presets

ShapeMetriX offers three reconstruction presets – each created with a specific balance of resolution and performance in mind: 

  • Fast preset: Generates low-resolution topography and texture. Suitable for quick or rough models, or bulk volume calculations where detail is less critical. 
  • Normal preset: The default option provides medium-resolution topography with high-resolution texture. Good for balanced resolution and processing speed. Typically used for mid-level geomatics tasks and applications like blast design, mapping single faces, benches, or basic rock mass characterization.  
  • High preset: Delivers high-resolution topography and texture for maximum detail and quality. Recommended for detailed rock mass characterization or when capturing fine surface details is essential. 

Use High Dense Reconstruction Preset and proceed to dense reconstruction process. 

Figure 4. Dense reconstruction settings.
3.2 Review Final 3D Model

Inspect the resulting 3D model in the 3D viewer.

Figure 5. Final 3D model.

Review the automatically generated Reconstruction Report, which provides a detailed summary of both coarse and dense reconstructions, along with model referencing and relevant statistics.

3.3 How Presets Influence 3D Model Quality 

Let’s go through steps 1 to 3 again, this time selecting a different Dense Reconstruction Preset each time. This will produce 3D models with varying mesh and texture quality and resolution for comparison purposes. 

Re-processing the same dataset under each preset demonstrates how model fidelity shifts from coarse topography to millimeter-scale detail. The summary statistics below quantify these trade-offs. 

Dense Reconstruction Preset  Dense Points Generated  Mesh Points  Mesh Triangles  Mean Point
Spacing (m) 
Fast  1,633,123  668,286  1,333,992  0.130 
Normal  7,656,542  2,916,982  5,828,181  0.064 
High  30,622,939  11,866,489  23,717,309  0.033 

Table 1. Comparison of Dense Reconstruction Presets: point and mesh statistics
 

Step 4. Compare Geometry, Mesh, and Texture Resolution 

4.1 Geometry and Mesh Resolution 

Higher-density presets generate a greater number of points and finer mesh triangles, capturing small features and subtle surface details while requiring more computational time. Lower-density presets result in coarser, less detailed models in a shorter processing time.

Figure 6. 3D reconstructed surface – fast dense reconstruction preset (top) and high dense reconstruction preset (bottom).
4.2 Texture Resolution 

Higher-density presets improve the texture resolution, closely replicating the original object’s shape and appearance. Coarser density presets lead to reduced resolution and less detailed topography and textures. 

Figure 7. Texture draped onto the irregular surface – fast dense reconstruction preset (top) and high dense reconstruction preset (bottom).

The Verdict 

In practice, the trade-off is straightforward: when precision in rock mass characterization or structural mapping is critical, the high-density preset delivers the textural and geometrical fidelity that detailed projects demand. When fast turnaround takes priority, lower-density presets keep workflows agile without sacrificing essential outcomes.

The finer surface meshes generated at high density are especially valuable when transferring models into Rocscience tools such as RocSlope3 or RocTunnel3, where realistic geometries and mapped discontinuities form a more reliable foundation for geotechnical analysis.

Key Takeaways 

From rapid volume estimates to detailed structural mapping, ShapeMetriX scales to the demands of each project. The choice of dense reconstruction preset is the control point: it defines model resolution, governs processing effort, and ensures outputs are both usable and efficient. By aligning preset selection with project goals, workflows stay streamlined and results remain fit for purpose.

3GSM dark logo
Simply Measure, Tools Redefined
Products
Sectors
Resources
Contact Us
Follow Us

Privacy   |   Legal Notice