4
et al. (2019) visualize ground settlement of the surface
above an underground tunnel utilizing analytical methods
of settlement with BIM models of the surrounding build-
ings on the surface, near the tunnel. Even though analytical
methods are used instead of monitoring data, simple yet
clear visualizations are considered an impactful decision-
making tool. A similar case study was conducted by Ninić
et al. (2024) where soil settlement above an underground
tunnel was analyzed utilizing a BIM (Building Information
Modeling) framework. Kim et al. (2020) used geostatisti-
cal spatial interpolation techniques to visualize and esti-
mate the distribution of soil properties over a surface using
borehole datasets. Their study focuses on technical visu-
alization techniques over a soil surface area and discusses
the predictions from these visualizations to the observed
values. Studies such as Kim et al. can be great applications
for other rock mass data sets for the underground mining
industry. Similar to this study, Lyu &Wang (2024) visual-
ize subsurface geology in 3D using borehole datasets using
a method called BCS. Civelekler &Pekkan (2022) develop
3D visualizations of soil properties by applying them in
GIS, which they believe can provide preliminary informa-
tion for scientists and engineers. Wang et al. (2020) develop
a geotechnical visualization tool for large-scale simulations
for mobile devices instead of desktops and laptops. The
studies mentioned above give great foundational work and
methodologies to create scientific visualizations and simu-
lations with impactful visualizations. However, they do not
incorporate rock mass data such as movement and strain.
Sekiya et al. (2022) tackle this gap in geotechnical visual-
ization by creating a 3D visualization of tunnel opening
deformation by acquiring data from MEMS accelerome-
ters. Similarly, Liang et al. (2024) developed a data-driven
3D model integrated with real-time data visualization. The
datasets are visualized over the corresponding spatial area
of the underground mine as simplistic glyphs or symbols,
where the underground space is represented with CAD
files. Their visualizations also include geological structures
in corresponding space at the mine. Similarly, Seymour et
al., (2016) visualized seismic data in 3D. The seismic data
was represented around the 3D model of the mine accord-
ing to its georeferenced space, in a 3D scene.
Keeping in mind the methods used for the examina-
tion of rock mass, and the literature, this paper aims to pro-
vide insight into how to develop a scientific visualization
of continuous rock mass displacement considering its rela-
tions to space on the rock mass over time. This visualization
method for spatiotemporal data for underground mines
utilizes quantitative data obtained by MPBX instrumenta-
tion as mentioned in Section 1.1. The methods (Section
4) for developing the visualization include utilizing 3D
meshes to represent the East decline, colors rendered over
the 3D mesh to represent the state of displacement, and
desired timelapse animations to illustrate near-real-time
representations. The goal of developing scientific visualiza-
tions of rock mass data for underground geomechanics is
to provide:
• A visual display of the data, representing the state of
the rock mass for safety and geotechnical analysis,
• A 3D visualization of the opening to represent the
position of an incoming datum, and its relations
with other data,
• A visual display of the rock movement phenomena
over time.
METHODOLOGY
The methodology consists of three major components:
(1) generating, editing, and implementing 3D meshes,
(2) ºdeveloping an algorithm to differentiate the areas of
an underground opening, and (3) developing an algorithm
to interpolate displacement data over two points across an
area on the 3D mesh of the underground opening. Cloud
Compare [https://www.danielgm.net/cc/], a point cloud
editing and analysis software, and Blender [https://www.
blender.org/], a 3D graphics software, were utilized to cre-
ate and edit 3D meshes of step (1). The developed meshes
then were used as the base to render the visualization algo-
rithms. Unity [https://unity.com/], a real-time 3D develop-
ment platform, was adopted to develop the final 3D scene
as it provided an environment to conduct steps (2) and (3).
Step (1) required manual acquisition and manual editing
of point clouds or the creation of a simplistic representa-
tive 3D model. Step (2) and step (3) were automated with
the script after step (1) was completed and imported into
Unity. Step (2) and (3) are conducted by shader programs
specifically, HLSL (High-level Shader Language) used in
Unity.
Step 1: Model Preparation for Unity
This step as mentioned involves 3D model generation and
editing. The edits include UV unwrapping and scaling of
the model. Mesh generation can be conducted in two ways
depending on geotechnical engineers’ preference for preci-
sion and time consumption. One way of generating 3D
models is by utilizing point cloud data from Lidar scans,
Laser scans, or photogrammetry scans of openings at the
East decline. After acquiring the point cloud data, they are
transformed into a 3D mesh made up of low triangulations
in Cloud Compare utilizing built in tools. Since the point
et al. (2019) visualize ground settlement of the surface
above an underground tunnel utilizing analytical methods
of settlement with BIM models of the surrounding build-
ings on the surface, near the tunnel. Even though analytical
methods are used instead of monitoring data, simple yet
clear visualizations are considered an impactful decision-
making tool. A similar case study was conducted by Ninić
et al. (2024) where soil settlement above an underground
tunnel was analyzed utilizing a BIM (Building Information
Modeling) framework. Kim et al. (2020) used geostatisti-
cal spatial interpolation techniques to visualize and esti-
mate the distribution of soil properties over a surface using
borehole datasets. Their study focuses on technical visu-
alization techniques over a soil surface area and discusses
the predictions from these visualizations to the observed
values. Studies such as Kim et al. can be great applications
for other rock mass data sets for the underground mining
industry. Similar to this study, Lyu &Wang (2024) visual-
ize subsurface geology in 3D using borehole datasets using
a method called BCS. Civelekler &Pekkan (2022) develop
3D visualizations of soil properties by applying them in
GIS, which they believe can provide preliminary informa-
tion for scientists and engineers. Wang et al. (2020) develop
a geotechnical visualization tool for large-scale simulations
for mobile devices instead of desktops and laptops. The
studies mentioned above give great foundational work and
methodologies to create scientific visualizations and simu-
lations with impactful visualizations. However, they do not
incorporate rock mass data such as movement and strain.
Sekiya et al. (2022) tackle this gap in geotechnical visual-
ization by creating a 3D visualization of tunnel opening
deformation by acquiring data from MEMS accelerome-
ters. Similarly, Liang et al. (2024) developed a data-driven
3D model integrated with real-time data visualization. The
datasets are visualized over the corresponding spatial area
of the underground mine as simplistic glyphs or symbols,
where the underground space is represented with CAD
files. Their visualizations also include geological structures
in corresponding space at the mine. Similarly, Seymour et
al., (2016) visualized seismic data in 3D. The seismic data
was represented around the 3D model of the mine accord-
ing to its georeferenced space, in a 3D scene.
Keeping in mind the methods used for the examina-
tion of rock mass, and the literature, this paper aims to pro-
vide insight into how to develop a scientific visualization
of continuous rock mass displacement considering its rela-
tions to space on the rock mass over time. This visualization
method for spatiotemporal data for underground mines
utilizes quantitative data obtained by MPBX instrumenta-
tion as mentioned in Section 1.1. The methods (Section
4) for developing the visualization include utilizing 3D
meshes to represent the East decline, colors rendered over
the 3D mesh to represent the state of displacement, and
desired timelapse animations to illustrate near-real-time
representations. The goal of developing scientific visualiza-
tions of rock mass data for underground geomechanics is
to provide:
• A visual display of the data, representing the state of
the rock mass for safety and geotechnical analysis,
• A 3D visualization of the opening to represent the
position of an incoming datum, and its relations
with other data,
• A visual display of the rock movement phenomena
over time.
METHODOLOGY
The methodology consists of three major components:
(1) generating, editing, and implementing 3D meshes,
(2) ºdeveloping an algorithm to differentiate the areas of
an underground opening, and (3) developing an algorithm
to interpolate displacement data over two points across an
area on the 3D mesh of the underground opening. Cloud
Compare [https://www.danielgm.net/cc/], a point cloud
editing and analysis software, and Blender [https://www.
blender.org/], a 3D graphics software, were utilized to cre-
ate and edit 3D meshes of step (1). The developed meshes
then were used as the base to render the visualization algo-
rithms. Unity [https://unity.com/], a real-time 3D develop-
ment platform, was adopted to develop the final 3D scene
as it provided an environment to conduct steps (2) and (3).
Step (1) required manual acquisition and manual editing
of point clouds or the creation of a simplistic representa-
tive 3D model. Step (2) and step (3) were automated with
the script after step (1) was completed and imported into
Unity. Step (2) and (3) are conducted by shader programs
specifically, HLSL (High-level Shader Language) used in
Unity.
Step 1: Model Preparation for Unity
This step as mentioned involves 3D model generation and
editing. The edits include UV unwrapping and scaling of
the model. Mesh generation can be conducted in two ways
depending on geotechnical engineers’ preference for preci-
sion and time consumption. One way of generating 3D
models is by utilizing point cloud data from Lidar scans,
Laser scans, or photogrammetry scans of openings at the
East decline. After acquiring the point cloud data, they are
transformed into a 3D mesh made up of low triangulations
in Cloud Compare utilizing built in tools. Since the point