Content uploaded by Emanuela Genovese
Author content
All content in this area was uploaded by Emanuela Genovese on Jul 01, 2024
Content may be subject to copyright.
Advances in Bamboo Science 7 (2024) 100079
Available online 18 April 2024
2773-1391/© 2024 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-
nc/4.0/).
Bamboo structures: Innovative methods and applications for structural
health monitoring and dissemination
☆
Vincenzo Barrile
*
, Emanuela Genovese
Department of Civil Engineering, Energy, Environment and Materials (DICEAM) - Mediterranea University of Reggio Calabria, Via Zehender (loc. Feo di Vito), Reggio
Calabria, 89124, Italy
ARTICLE INFO
Keywords:
Bamboo
Bamboo construction
Structure for motion
YOLO
Neural networks
Virtual/Augmented/Mixed reality
ABSTRACT
Bamboo is gaining more and more attention in the eld of building materials thanks to its unique characteristics,
such as resistance, lightness, sustainability, and exibility. However, the widespread adoption of bamboo as a
building material presents some challenges, especially in relation to the durability of the material. In fact, being a
natural material, it requires protection from atmospheric agents, such as rain and sun, as well as from attacks by
insects and animals. Ensuring the longevity of bamboo involves implementing tailored treatments, precise
precautions, and regular structural inspections to detect any potential degradation over time. In this context, the
article focused on the monitoring of a structure made of bamboo and on the testing of an innovative system to
identify the structural critical issues present, in the context of a FISR project which envisaged a cultural exchange
between China and Italy. Starting from the survey of the structure carried out with a drone, a 3D model was then
built using a commercial software. For the monitoring phase of the structure, an innovative system was used that
exploits the YOLO v5s6 algorithm to identify structural critical issues, supported by a Virtual/Augmented/Mixed
reality app developed by the authors. This app is useful not only during the monitoring phase, but also in the
subsequent phase of disseminating information relating to the potential of the use of bamboo in the eld of
structural engineering. The tested and proposed methodologies have proven to be particularly useful and high-
performance, especially regarding the monitoring of bamboo structures, in relation to the resolution of problems
linked to their possible fragility and limitations in certain application areas. The contribution provided in this
paper by Geomatics methodologies is evident to better highlight the potential of this eco-friendly material for its
desirable ever-increasing use in the construction sector by becoming aware of its potential.
1. Introduction
Bamboo is a plant that is part of the Graminaceae (or Poaceae) family
and the Bambusoideae subfamily. It is a woody herbaceous plant,
characterized by hollow and knotty stems called culms. It is a material
known for its rapid growth and resistance and it is widely distributed in
many parts of the world, especially in tropical and subtropical regions,
(Cheng et al., 2023; Ben-Zhi et al., 2005). Globally, there are approxi-
mately 1250–1500 species of bamboo, comprising approximately
75–107 genera, (Kumar et al., 2023; Han et al., 2023).
From an environmental point of view, bamboo is attracting more and
more interest globally for its sustainability. The research develops on
two main fronts: on the one hand, it is studied and appreciated as a
sustainable material that does not require complex manufacturing
processes; on the other hand, it is engineered (Janssen et al., 1981)
through the lamination of bamboo or LBT (Laminated Bamboo Timber),
which involves dividing the material into strips and recomposing them
through the use of glues (Sharma et al., 2015; Shu et al., 2020). This
technique allows to overcome the natural geometric limitations of
bamboo, although it could be considered a loss of some of its intrinsic
structural characteristics. There are many advantages of this natural
material; in fact, it grows quickly, requires few resources for its culti-
vation and has a low environmental impact compared to other tradi-
tional building materials such as wood or steel. This makes it an
attractive option for sustainable construction and carbon emissions
reduction. In fact, there are several studies that aim to quantify the
carbon emissions of bamboo buildings and analyze their potential for
reducing carbon emissions (Zhu et al., 2023). Xu et al., 2022 differs from
☆
This article is part of a special issue entitled: “Special issue on Future uses of bamboo” published at the journal Advances in Bamboo Science.
* Corresponding author.
E-mail address: vincenzo.barrile@unirc.it (V. Barrile).
Contents lists available at ScienceDirect
Advances in Bamboo Science
journal homepage: www.journals.elsevier.com/advances-in-bamboo-science
https://doi.org/10.1016/j.bamboo.2024.100079
Advances in Bamboo Science 7 (2024) 100079
2
previous studies because it explores the carbon emissions and storage of
bamboo building materials by considering each phase (planting phase,
transportation phase, production phase and construction phase) and
where an overview of carbon emissions and storage of bamboo assem-
bled components is presented.
Bamboo is a renewable natural resource characterized by a unique
brous structure that imparts remarkable strength and exibility to the
material. Its tensile strength is comparable to that of steel, while its
lightweight nature makes it an ideal choice for a wide range of structural
applications. However, to fully harness the potential of bamboo in
construction, it is essential to understand its mechanical properties in
detail.
From a structural point of view, bamboo has various characteristics
that make it unique in its kind. These mechanical characteristics are
inuenced by its structural composition and there are many studies
aimed at determining the parameters that inuence them (Emamver-
dian et al., 2020; Wang et al., 2020; Lou et al., 2021).
As it is known, the specic characteristics of bamboo can vary
depending on the species (Escamilla et al., 2018; Fahim et al., 2022), age
and growth conditions, but those common to all types of bamboo can be
summarized in resistance, lightness and exibility.
Particular attention from an engineering point of view is being paid
to the so-called Glubam. Glubam is produced by joining thin layers of
bamboo together using adhesives and pressure to form a strong and
durable composite material. The process of making Glubam involves the
selection and preparation of bamboo, which is cut into thin strips and
then glued together. The layers of bamboo can be arranged in different
directions, like the thread of wood, to achieve the desired mechanical
properties. The resulting material is then subjected to pressure and heat
to ensure good adhesion between the layers and greater solidity of the
structure. Glubam offers several advantages as a building material. First,
it has excellent mechanical properties, including tensile, compressive,
and exural strength, making it suitable for a variety of structural ap-
plications. Additionally, this material is known for its dimensional sta-
bility, meaning it is less prone to warping than solid bamboo. This makes
it a reliable option for use in buildings and structures. Another advan-
tage of Glubam is its greater resistance to weather, insects and animals
compared to untreated bamboo. Thanks to the use of adhesives and the
lamination process, Glubam is more resistant to humidity and degra-
dation, thus increasing its durability over time. The use of Glubam as a
construction material offers several advantages, primarily due to its
ability to reduce overall environmental impact compared to conven-
tional building materials. Bamboo is a renewable resource that requires
fewer resources and less energy to produce, resulting in a lower carbon
footprint. The utilization of Glubam in construction also promotes sus-
tainable forestry practices. Bamboo grows rapidly and can be harvested
without causing long-term damage to the environment. Additionally, its
strong and versatile nature makes it suitable for a wide range of con-
struction applications, further enhancing its appeal as an environmen-
tally friendly alternative. Therefore, it enables a greener and more
environmentally conscious approach to construction.
Although bamboo is a natural material, it can boast a set of regula-
tory documents that legitimize and regulate its use. Unlike other natural
materials, bamboo has gained widespread regulatory acceptance. In
some countries such as Colombia, Peru, Ecuador and India, bamboo is
included in national legislation with the same rights as materials
commonly used in construction. In other parts of the world, there are
standardization rules that establish criteria for the structural use of
bamboo. Beginning in 2000, the ICBO Evaluation Service, Inc., under
the International Code Council, introduced evaluation standards for
bamboo in the United States. In 2004, ISO standards were established for
bamboo in Europe. These standards provide guidelines for the design,
construction, and safe use of bamboo in structures. The presence of
regulatory documents specic to bamboo indicates recognition of its
validity as a building material and provides a basis for the widespread
and safe use of bamboo in architectural and construction practices. This
helps promote the adoption of bamboo as a sustainable and versatile
resource in the construction industry.
The possibility of using bamboo as a primary material in the con-
struction industry is given by the resistance capacity of the material. As
known, it develops over time, reaching its maximum between 3 and 5
years of growth, in conjunction with the process of formation and
hardening of cellular tissues. It is important to note that the resistance
varies depending on the different components of the bamboo, inu-
encing the response to external agents in diverse ways. In areas where
there is a greater concentration of vessels and bers, the specic weight
increase in a manner correlated to the performance of the material: the
greater the density and specic weight, the greater the resistance. For
example, in general, the upper part of a bamboo culm will have higher
mechanical performance than its lower part, and the same applies to the
external part of the section compared to the internal one. However, this
is not always true. From the literature, it emerges that if bamboo has
been subjected to unfavorable environmental conditions or structural
damage, there may be a variation in the distribution of mechanical
properties along the culm. Additionally, factors such as bamboo species,
age of the plant, and growth process can inuence the mechanical
characteristics of the material. Another principal factor is the moisture
content as reported by Sylvayanti, et al. (2023): “Moisture content is
crucial, especially on hygroscopic materials, because it affects their
volume and mass, affecting other properties.” The compressive strength
of bamboo varies depending on the direction in which the load is
applied. When the load is parallel to the orientation of the bers,
resistance is maximized, while when the load is orthogonal, resistance is
minimized. Furthermore, bamboo’s behavior varies depending on where
the load is applied along the culm, with greater performance in the
internode areas., (Greco, 2017).
In other words, as it can be seen from the different bibliography
present in international literature, bamboo shows a notable ability to
resist tensile and bending forces with resistance values comparable to
those of steel and which indicate its suitability as a structural element.
These results highlight the exceptional mechanical properties of bamboo
as a structural material, (Hegde and Sitharam, 2015; Yadav and Mathur,
2021).
For this reason, in recent years, bamboo has gained new prestige in
the eld of architecture at an international level. Another factor that
favors the widespread use of bamboo is its adaptability to emergency
situations, such as in the case of natural disasters or immediate housing
needs. Its affordability, availability, simplicity, and speed of use, com-
bined with its high mechanical performance, make it an excellent and
frequent choice in these situations, especially in Asian countries.
As regards shear strength, the literature reports average shear
strength values of around 12 MPa for (Greek) bamboo. These relatively
low values must be considered in light of the structure of the material
itself, which has no transversal bers but only a cohesive matrix that is
unable to resist signicant stress. The joints between the elements
represent a weak point in the design of bamboo structures. This is due to
several reasons, including:
1. Differences between the elements to be connected: Bamboo is a
natural material, and the elements used may present variations in
diameter, thickness, and section. These differences can affect the
strength and stability of the joints.
2. Circular and hollow prole: It has a circular prole and a hollow
section, which makes the connection between the elements more com-
plex than materials with more regular sections, such as wood or steel.
3. Rigid and smooth outer surface: The outer surface of bamboo is
stiff and smooth, which can make it difcult to achieve good adhesion
and tensile strength in joints.
4. Poor shear strength: It has poor shear strength due to its composite
structure, which consists primarily of a cohesive matrix without signif-
icant transverse bers. This can limit the material’s ability to withstand
shear stresses in joints.
5. Orthotropicity: Bamboo is an orthotropic material, which means
V. Barrile and E. Genovese
Advances in Bamboo Science 7 (2024) 100079
3
that its mechanical behavior varies depending on the direction of the
applied stresses. Since bamboo bers grow only in the longitudinal di-
rection, the material is inhomogeneous in its response to stresses applied
parallel or orthogonal to the surface of the element. This can result in
uneven stress distribution in the joints. However, when it comes to di-
aphragms, there might be a greater dispersion of bers, thus leading to
variation in mechanical properties compared to parts of the culm with
more directional bers.
Therefore, when designing bamboo structures, it is crucial to address
these joint challenges and adopt appropriate solutions to ensure the
overall strength, stability and durability of the structure.
To address these issues, several connection, and jointing techniques
specic to bamboo have been developed. These techniques include the
use of nails, screws, bamboo bindings, metal connectors and the use of
resins or glues (Moran and García, 2019; Widyowijatnoko and Harries,
2020; Ghavami and Moreira, 1996; Seixas et al., 2021).
In addition to the connection problems between the various ele-
ments, as mentioned previously, bamboo presents problems related to
durability. The bamboo used in structures, in fact, can present various
degradation problems which can compromise its resistance and dura-
bility over time. Bamboo is a porous material that can absorb moisture
from its surroundings, encouraging the growth of mold, mildew, and rot,
which can weaken the structure. Another widespread problem is attacks
by insects and parasites. Bamboo is susceptible to attacks from termites,
ants, and beetle larvae, which can burrow into the material, causing
considerable damage to the structure (Yadav and Mathur, 2021). Nat-
ural degradation is another factor to consider. Over time, bamboo can
undergo a process of aging and degradation linked to exposure to at-
mospheric agents. This can lead to a loss of strength and the formation of
cracks or fractures. Finally, movement and deformations of bamboo can
be a problem. This material is subject to movement and deformation,
especially in response to changes in humidity and temperature. Such
movements can cause cracking or detachment of joints in structures. To
prevent and address these issues, it is essential to adopt prevention and
protection measures. This may include the application of pesticide and
antifungal treatments, the use of protective nishes to reduce moisture
absorption, as well as the use of robust joints and connections to ensure
structural stability (Zhang et al., 2023). A full comprehensive review by
Bala and Gupta, 2023 reports not only the physical and mechanical
properties of bamboo structures and a detailed description of the re
performance of bamboo culm and engineered bamboo but also sum-
maries the principal key issues of bamboo elements and their solutions:
“Round culm bamboo is not suitable for structural applications whereas,
laminated bamboo and bamboo scrimber are value added EBPs, most
popular for structural applications due to their standard shape and size
with controlled variation in physical and mechanical properties along
the longitudinal, radial and tangential directions.” Furthermore, it is
important to conduct regular checks and maintenance to promptly
identify signs of deterioration and intervene accordingly.
In this context, the ability to manage information regarding the state
of degradation of bamboo and in particular of the joints in constructions
made with this material is undoubtedly useful.
In relation to the survey and monitoring phases of structures and
infrastructures in general, there is a growing interest in the use of UAVs
(Remotely Piloted Aerial Vehicles). These technologies allow to carry
out detailed surveys, becoming a versatile tool to identify problems at a
structural level, such as cracks, lesions or any type of damage to the
building or structure. In the literature, it is possible to nd several
studies that deal with the use of drones for structural surveying. UAVs, in
fact, are frequently employed to construct 3D models and for inspections
especially regarding reinforced concrete structures and infrastructures
(Ivi´
c,et al., 2023); however, there is limited literature on their use in
monitoring the condition of a structure and conducting quantitative
analysis with Structure from Motion techniques, less can be found
regarding the use of this technology to monitor bamboo structures
(Chaudhry, et al., 2020; Heichel et al., 2023)
The process through which it is possible to conduct these types of
investigations is based, as is known, on Structure for Motion technology.
It involves the use of the images captured by the drone to create a point
cloud (rst sparse and then dense) and, consequently, a 3D model, using
conventional photogrammetric techniques. The photogrammetric
dataset, in fact, rather than coming from an airplane, comes from a
drone, which, following a pre-set ight plan, depending on certain pa-
rameters, takes a certain quantity of images which, once processed using
appropriate image processing software, provide as output the 3D model
of the detected structure or infrastructure, which is measurable and
scalable (Barrile et al., 2021; Candela et al., 2019). Within the panorama
of photogrammetric software, it is possible to choose between algo-
rithms and programs with distinct functions and methodologies, each
optimized for certain types of survey or instrumentation used. In this
study, we focused on the Agisoft Metashape software, based on the
Structure for Motion algorithm, which allowed to reconstruct a point
cloud using photographs and from this a measurable and usable
three-dimensional mesh for the purposes of the study (Kleinsmann et al.,
2023; Godone et al., 2020).
Regardless of the use of the images acquired by the drone for the
construction of the 3D model, they can still be processed subsequently or
in real-time with Machine Learning techniques which, using sophisti-
cated algorithms, allow the identication of particular regions of in-
terest. There are several types of algorithms depending on the objectives
and desired results, and numerous studies highlight their advantages
and disadvantages. In the context of Structural Health Monitoring,
articial intelligence and machine learning are gaining more and more
importance (Gupta, 2023; Zhong et al., 2022).
In fact, machine learning algorithms can analyze enormous amounts
of data and predict potential structural problems with high precision. In
general, they can be used for structural monitoring, highlighting struc-
tural variations or anomalies, and identifying potential problems. These
algorithms allow the prediction of structural deterioration, (in some
cases particularly relevant for bamboo structures), using historical data
and other parameters characteristic of the environment and the struc-
ture under examination. Identifying structural problems is, in fact, an
area where the application of machine learning algorithms can be
extremely useful. For example, Zhou et al., 2021 used Support Vector
Machines to develop an automated method to translate hysteresis loop
analysis results obtained from structural health monitoring into
nonlinear fundamental models. Other studies, such as Chang et al.,
2018, have employed neural networks to interpret damaged structures
in terms of damage locations and severity, as well as the residual per-
formance of damaged elements. In another study, Gu et al., 2021, a
method based on evidence theory and a random forest algorithm was
developed to extract the factors inuencing a dam. There are also al-
gorithms capable of detecting objects in real time in images and videos,
such as YOLO (You Look Only Once) (Lan et al., 2018; Liu et al., 2018;
Chandana and Ramachandra, 2022).
Although its original application was designed for the purpose of
identifying people or objects in the frames of a video or image, YOLO
algorithm can be used to identify objects or patterns in complex struc-
tures. This algorithm uses a convolutional neural network architecture
to divide the image into a grid and create "bounding boxes" in which the
object ts within a cell of the grid (PyTorch, 2023).
The information acquired through the analysis of the structure, as
well as the characteristics of the material themselves, need to be shared
with an increasingly wider audience (both for the personnel responsible
for managing the infrastructure and for the simple user who wants to
know its potential), the concept of digitalization of information is
becoming increasingly established. To this end, the information ac-
quired relating to a 3D model and more generally that relating to the
state of deterioration of structures and infrastructures can be profes-
sionally managed using suitable dedicated virtual/augmented and
mixed reality apps. Such applications are widely known in this eld
(Chiang et al., 2022; Arena et al., 2022; Wang et al., 2022) and allow the
V. Barrile and E. Genovese
Advances in Bamboo Science 7 (2024) 100079
4
acquired information to be visualized, analyzed, and disseminated. In
fact, if on the one hand these apps allow to analyze the data collected for
monitoring structures, on the other hand they can represent powerful
tools for informing on the potential of still little-known materials, such
as bamboo (Kohli et al., 2022; Hasan et al., 2022; Martins et al., 2022).
In this context, the present study addressed the survey of a structure
whose roof is made of bamboo as part of the FISR/GoForIT/CRUI
Foundation Project “Seismic assessment of bridges and viaducts with
Remotely Piloted Aircraft Systems (RPAS) and Articial Vision” in
collaboration with Zhejiang University (China) and the Department of
Engineering Structure of the Sapienza University of Rome. The CRUI
Foundation made a research grant available to the Mediterranean Uni-
versity of Reggio Calabria which allowed a PhD from the Mediterranean
University to gain professional experience in a foreign institution,
namely Zhejiang University in China.
Initially, the survey was carried out using a DJI Mavic 2 Pro drone
and subsequently the images were processed using a commercial soft-
ware called Agisoft Metashape, which uses typical Structure from Mo-
tion algorithms. This process allowed to create a three-dimensional
model of the structure in a detailed and accurate manner. Subsequently,
through the YOLO v5s6 algorithm, an automated and innovative method
based on a deep learning object detection model (YOLOv5s6) was
implemented to capture and identify deteriorations in bamboo struc-
tures using bounding boxes which allowed the identication of prob-
lems relating to joints of the selected bamboo structure. The authors
then concentrated on the implementation of a virtual, augmented and
mixed reality application, which made it possible to analyze the
different parts of this structure as well as disseminate the information,
thus making it better available and usable by both the operators assigned
to maintenance and to the individual user in order to know the prop-
erties of these types of structures. This app is particularly useful during
the monitoring and management phase of the structure, and it has some
advantages. The application, in fact, allows to view the structure in an
immersive and interactive way, allowing operators to explore details
and specic parts of the structure without having to be physically pre-
sent on the site. This reduces the need for travel and physical access to
potentially dangerous or difcult to reach areas, increasing operator
safety. Furthermore, the application can provide additional information
on the state of degradation of the structure, through the integration of
data obtained from the use of YOLO. This allows a more accurate
assessment of the structural conditions and facilitates the planning of
maintenance or intervention activities. Finally, it can promote
communication by raising awareness about the use of this innovative
and eco-friendly material. This therefore allows an innovative and dual
purpose of monitoring but also of disseminating the opportunities that
this material can bring.
2. Materials and methods
The methodology advanced in the following study was applied to a
structure located at the entrance of the Zhejiang University, Yuhang-
tang, Hangzhou, located in the eastern part of China, a higher education
institution that hosts the Faculty of Civil Engineering (Civil Engineering
College) in relation to FISR/GoForIT/CRUI Foundation Project “Seismic
assessment of bridges and viaducts with Remotely Piloted Aircraft Sys-
tems (RPAS) and Articial Vision” in collaboration with Zhejiang Uni-
versity (China) and the Department of Engineering Structure of the
Sapienza University of Rome. Fig. 1 shows a map of the area in which the
detected bamboo structure is located.
The roof system structure, erected in 2017, consists of a space beam
composed of Glubam and steel, with dimensions of 12 m x 3.6 m. This
structure is composed of 2 ×8 identical square pyramids, with the vertex
positioned on the lower curb. The base of each module is
1200 mm×1200 mm, while the height is 849 mm. Accordingly, the in-
plane dimensions of this space structure are 2400 mm×9600 mm.
In order to develop a virtual/augmented/mixed reality application
with the dual purpose of visualizing and monitoring the state of damage
and deterioration of structures, as well as disseminating information on
this eco-sustainable material and its engineering applications, it is
necessary to follow a series of steps: the survey, the 3D modeling of the
structure, the application of the YOLO v5s6 model to the images of the
structure, and the import of the model and information relating to the
structure within the virtual/augmented/mixed reality application.
As regards the survey of the structure, we proceeded using the Mavic
2 Pro drone. This drone uses a high-resolution camera with integrated
GPS to geotag each photo and record the location information in EXIF
les. The main characteristics of the drone mentioned above are shown
in Table 1:
The ight plan for this study was programmed in such a way as to
obtain a Ground Sampling Distance of 0.5 cm/pixel guaranteeing a
horizontal overlap of 80% and a vertical overlap of 20%. To set the ight
plan, the commercial application Pix4D was used in free ight mode,
taking photos on a 1 m x 1 m grid both vertically and horizontally. The
reason we chose to use drones for the survey, along with the YOLOv5s6
algorithm, stems from several key advantages they offer over available
alternatives. Beginning with drones: their exibility and accessibility are
among the primary benets. Thanks to their ability to reach areas
difcult to access with traditional methods, such as rugged terrain or
hazardous environments, drones ensure comprehensive coverage of the
survey area. Additionally, due to the high spatial resolution of the im-
ages captured by drones, we can obtain precise details to analyze and
map terrain features accurately. Economically, drones are a cost-
effective choice and provide near-real-time data, allowing for rapid
decisions and timely responses to any changes in the survey environ-
ment. Turning to the YOLOv5s6 algorithm, it was chosen for several
reasons. Firstly, it offers state-of-the-art performance in real-time object
detection. Its speed and accuracy are extraordinary, making it ideal for
integration with drones, where limited processing capability requires
efcient algorithms. Its excellent precision and generalization capabil-
ities make it suitable for a wide range of object classes and
Fig. 1. Cartographic representation (extrapolated with the QGIS software via
Google Labels and Google Satellite basemaps) of the area in which the bamboo
structure is present at Zhejiang University, China.
Table 1
Characteristics of the Mavic 2 Pro drone used for the survey phase.
Sensor Sony FC220 – CMOS 1/2.3”
Image resolution 12.35 megapixels
Lens 28 mm f/2.2
Real focal length 5 mm
Effective sensor width 6.7 mm
ISO range 100–1600
Electronic shutter speed 8 s – 1/8000 s
Image size 4000 ×3000 pixels
Geotagging Built-in GPS
V. Barrile and E. Genovese
Advances in Bamboo Science 7 (2024) 100079
5
environmental conditions. Additionally, the ease of deployment and
integration of YOLOv5s6 into existing workows is another signicant
advantage. Thanks to its open-source nature and detailed documenta-
tion, it can be customized and adapted to specic use cases.
The image processing phase involves several steps: cleaning the
images to remove noise and artifacts using Gaussian blur; color correc-
tion, which adjusts brightness, contrast, and color balance; and pre-
processing, the most fundamental step, where the image is prepared for
input into the YOLO model. This involves standardizing the size and
normalizing pixel values. For training the YOLO model, datasets spe-
cically collected for the task and sourced from the web were chosen,
and data augmentation was performed. The YOLO variant was cong-
ured in its architecture by selecting the number of layers (200 layers)
and anchor boxes. In this particular case study, a moderate learning rate
was chosen since the focus was on bolt detection. Since inferences were
conducted on batches of images, a Batch Size of 32 (i.e., Speed V100
b32) was determined to be the most effective. As is well-known, "batch
size" is a parameter indicating the number of samples (e.g., images)
processed together in a single pass through the neural network during
training or inference. In other words, during model execution on a
dataset, the "batch size" denotes how many examples are processed
simultaneously. Subsequently, the model was trained using the prepared
datasets.
The structure of which the survey was carried out is characterized by
the presence of metal bolts that connect the various parts of the struc-
ture. Bolted joints are common elements of many engineering structures
which due to extreme service conditions and load factors can often be
subject to deterioration or loosening. Real-time monitoring of the
structure can be particularly useful in identifying loose bolts to ensure
the safety and durability of structures. As mentioned above, in the
literature it is possible to nd several studies in relation to the appli-
cation of machine learning and deep learning methods for the identi-
cation of deteriorated or loose bolts. However, most studies involve the
use of these algorithms in laboratory conditions, i.e. optimal conditions
with well-controlled light, distance and angle conditions. It is clear that
this practice is not practicable in large-scale monitoring because there
are many factors that inuence the images taken by the drone. For this
reason, in the present study, we chose to train the neural network model
via images taken by the drone.
This aspect, together with the combination of the YOLO algorithm to
identify deteriorations and loosening and the application which allows
simple and intuitive use of the information, both for trained personnel
and for the common citizen, constitutes the distinctive feature of this
study.
Following the image acquisition phase with the drone, a process of
image cleaning, color correction and segmentation of the affected areas
was therefore carried out using the KNIME software. Within this soft-
ware it is possible to insert the KNIME Image Processing extension which
allows the processing of images within the work environment, (Alha-
jahmad et al., 2023). The primary features of the KNIME Image Pro-
cessing extension are manipulation, analysis, ltering and visualization.
Once this treatment process was conducted, we proceeded with the
structural analysis for different deterioration scenarios by training the
neural network also using datasets freely available on the web (Fotia
et al., 2022). Its integration capabilities allow seamless merging of
image datasets with other data types, enabling holistic analysis. Through
Knime, users can perform transformative tasks like resizing, cropping,
and rotating images, ensuring uniformity across datasets. Moreover,
Knime facilitates sophisticated image enhancement techniques such as
contrast adjustment and noise reduction, crucial for improving image
quality. Additionally, Knime’s ability to normalize pixel intensities and
support batch processing ensures consistency and scalability in pre-
processing tasks. As for YOLO, there are several models available. The
parameters for evaluating the best model (Terven and Cordova-Esparza,
2023; Jiang et al., 2022; Majumder and Wilmot, 2023) are based on:
- size in pixels of the model which inuences precision and speed
- Mean Average Precision (mAP) evaluated on a scale ranging from
50% to 95% which represents the overall accuracy of the model in object
detection
- Mean Average Precision (mAP) assessed at a threshold of 50%.
- CPU execution speed in milliseconds (ms) for a single input image
- Execution speed on CPU (in this specic case V100) in milliseconds
(ms) for a single input image.
- Execution speed on CPU (in this specic case V100) in milliseconds
(ms) for a batch of 32 input images.
- The number of model parameters in millions.
- Number of oating-point operations per second (FLOPs) in billions
(G) during inference on an image with size 640 ×640 pixels.
Despite its effectiveness, we encountered several challenges during
testing. One major challenge was the variability in lighting conditions,
which affected the algorithm’s performance. In some cases, shadows or
reections could obscure critical structural features, leading to false
negatives or inaccurate detections. Addressing this challenge required
extensive preprocessing of the images to enhance contrast and minimize
the impact of lighting variations, partially addressed with the KNIME
software. Another challenge was the presence of occlusions, where
structural anomalies were partially obscured by other objects or back-
ground clutter. This posed a challenge for the algorithm, as it sometimes
struggled to distinguish between the structural anomaly and surround-
ing elements. To overcome this challenge, we experimented with
different augmentation techniques and ne-tuned the algorithm pa-
rameters to improve its robustness against occlusions. Moreover,
obtaining enough annotated data for training the model can be difcult,
especially for rare or specialized types of bolt deterioration. Augmen-
tation techniques may be necessary to articially expand the dataset.
Once the results of this analysis had been obtained and the struc-
turally deteriorated elements had been identied, it was possible to view
the information acquired (3D model, point cloud, images of deteriora-
tion) in an app appropriately created by the authors. The innovation of
this app consists in allowing to view information relating to a structure
or infrastructure in a fast and complete as well as interactive way,
allowing to have access to fundamental structural and construction in-
formation. Furthermore, by inserting the complete descriptive/technical
sheet, which can be consulted by the user, it is possible to analyze the
properties as well as the physical and structural characteristics of the
material used, in this case bamboo. This perspective has important ad-
vantages. Users can analyze the Bamboo structure with an in-depth level
of detail allowing a close-up view of its components, including the joints
used to create the structure. Detailed visualization through the app can
reduce inspection times in the eld and allows for more in-depth anal-
ysis without the need to be physically present. Furthermore, the appli-
cation could allow early visualization of signs of deterioration in the
structure, allowing preventive interventions before the damage becomes
serious. This can therefore be particularly useful in training any opera-
tors on the detection and management of deterioration. VR offers, in
fact, an immersive experience in which users can explore the structure
from different perspectives and provides detailed documentation that
can be archived for future reference and evaluation purposes. Further-
more, users can collaborate from different geographical locations by
discussing any interventions to be carried out in real time via shared MR.
Through multiple shots taken with the drone it is possible to create
multiple datasets useful for realizing a temporal evolution of the struc-
ture under examination.
In the panorama of possible software for creating virtual, augmented
and mixed reality applications, Unity 3D platform stands out among all.
This is a development environment that is often used in the eld of these
applications. It uses the integration of two types of code for its devel-
opment: Javascript and C#. Within this environment, scripts are called
“behaviors” and are used to capture assets in scenes and make them
interactive. Furthermore, GameObjects were used which represent
fundamental objects such as characters and props, equipped with
graphic representation, which were associated with Mono-Behavior
V. Barrile and E. Genovese
Advances in Bamboo Science 7 (2024) 100079
6
scripts to increase their functionality (Moiseienko et al., 2023; Jitendra
et al., 2021; Wang et al., 2021; Juliani et al., 2018).
The app developed allowed the digitization of the information ac-
quired, presenting itself as a catalog of 3D models and images that were
easily accessible and consultable, also for maintenance operators for
possible interventions. The app was designed to display the hologram of
the structure’s survey phases when the device frames the building, both
on site and remotely. In detail, the application allows to view the ho-
logram by aligning the coordinates of the device with those of the
structure within a radius of 10 km from it. Alternatively, you can view
the facility model and related information remotely. Mixed reality was
made possible thanks to the use of the Microsoft HoloLens tool with an
extension called Mixed Reality Toolkit through which the connection
between the virtual and augmented reality app and the HoloLens tool
was made possible (Karthika et al., 2017). Through HoloLens 2, it was
then possible to allow multiple users to interact in the same mixed re-
ality scene, allowing collaboration in virtual environments (Park et al.,
2021). The app displays the 3D model of the structure on the map,
associated with multimedia content, structural analysis, and links to
other nearby structures (Fotia and Barrile, 2023). Additionally, choosing
whether to enable GPS causes additional models to be displayed within a
10 km radius of the user. This application, in fact, uses external
augmented reality which uses GPS (but also the device’s accelerometer
and compass) to identify the position of the device with high precision; it
does not require special markers and uses technologies such as GPS to
position virtual objects in the real world.
For the implementation of Microsoft HoloLens, we proceeded
through various phases and commands, reported below:
1. Set up unity Environment:
- Unity Standard Conguration and Project Settings.
- Check universal Windows Platform Settings.
- Mixed Reality toolkit and application of Mixed Reality Project
Settings.
2.Structure Settings:
- Addition of Open Scenes, Select Universal Window Platform and
Selection of Target Device.
- Selection of Unity C# Projects and enablement of C#.
- Application of Mixed Reality Scene Settings and Starting without
debugging from Drop-down menu.
Once the developed environment has been dened, we proceed with
the 3D model import, associating the scripts and the basic settings to
ensure interaction with it through tags and labels. The Mixed Reality
Toolkit includes other scripts and tools to enhance the MR experience.
3. Results
During the survey phase of the structure, 459 photos were taken (for
a total of 2.5 GB) with a total time spent for the survey of 30 minutes.
Once a photogrammetric dataset suitable for the creation of the 3D
model was obtained, the image processing process was started using the
Agisoft Metashape software, after having previously used the KNIME
software for image pre-processing. Using this software, the 3D point
cloud and 3D mesh were generated separately. The generated point
cloud contained 3 million points and took 6 hours to complete. Fig. 2
shows the points cloud once processing is nished. The mesh consisted
of 952,634 points and 1867,241 triangles and took 1.2 hours to com-
plete. Fig. 3 4. shows a detail of the completed 3D model.
Once the 3D model was obtained, it was simplied using the 3D
graphics software Rhinoceros 6 and the Mesh2Surface plugin trans-
formed the acquired mesh into a CAD model.
From this 3D model, it was subsequently possible to extrapolate the
geometric characteristics of the structure, shown in Fig. 5.
Once the 3D model was created to be imported into the application
created by the authors, we proceeded with the analysis of the joints of
the structure under examination. These joints are made of metal, and we
wanted to check whether they could have deteriorated or loosened
following signicant loads and atmospheric agents. Through successive
temporal shots of the drone, it is possible, in fact, to make important
comparisons of these connections through the comparison between two
different frames acquired at two successive time instants t and t+1. The
possibility of easily comparing frames acquired at different moments is
made possible thanks to the use of UAVs by designing the acquisition
points through a suitable ight planner, with the aim of keeping the
distance and angle of the camera constant. The frame obtained at time t
is processed using an edge detection algorithm of the KNIME software to
improve the edges and facilitate comparison with the frame acquired
and processed at time t+1 (Fig. 6). In Fig. 6 the white line outlines the
edge of the bolt at time t while the red outline outlines the edge of the
bolt at time t+1. Obviously, the analysis was conducted on all the bolts
and any signicant deviation identied between the delimitation lines
of the same bolt at different times suggests the need for a more detailed
analysis on the potentially affected bolt.
For this reason, the images were then processed with the YOLO v5s6
algorithm (appropriately trained with datasets of images of bolts with
the characteristics of interest) to identify potentially deteriorated bolts.
In this case, there are two reference categories applied to the detection of
Fig. 2. Point cloud of the Bamboo structure after the 6-hour processing phase.
Fig. 3. Detail of the 3D model created using the Agisoft Metashape software.
Fig. 4. CAD model of the structure obtained from the survey conducted
by UAV.
Fig. 5. Geometric characteristics of the Glubam structure of the entrance of
Zhejinag University, China.
V. Barrile and E. Genovese
Advances in Bamboo Science 7 (2024) 100079
7
objects: "deteriorated" and "not deteriorated". Fig. 7 highlights the re-
sults obtained on the structure under investigation through the identi-
cation of the selected bolts identiable via the yellow bounding boxes.
This processing is sufcient to identify deteriorations, such as slips and
breakages. Sometimes identication can be difcult due to overlapping
portions of frames; therefore, to identify the bolts in these portions,
multiple frames are used which, using homologous points, allow their
identication.
Finally, for the main purpose of creating an information collection
tool for the possible planning of maintenance interventions by the
managing bodies and for the possible dissemination of information
regarding both the structure and the characteristics of the materials used
(bamboo), it was created a VR/AR/MR app as described in the Materials
and Methods section. Fig. 8(a) shows the hologram of the 3D model of
the structure when the device (with GPS activated) is near the structure.
Fig. 8(b), instead, shows the interaction of the 3D model remotely from
the Geomatics Laboratory of Mediterranea University of Reggio Cala-
bria. From the drop-down menu it is then possible to view other useful
sheets, which can be consulted both on site and remotely, such as the
technical sheet of the materials or a BIM model of the structure which
the user can share with other users simultaneously.
The potential of such an application is therefore evident both in
terms of information management in a simple and fast manner by the
managing body and in terms of engineering valorization of this material
whose mechanical and physical characteristics go well with the sus-
tainable nature of this material.
4. Discussion
The present study focused on the survey and evaluation of a bamboo
roof structure located in China. The approach adopted by the authors
involved the use of a DJI Mavic 2 Pro drone for the initial survey and
image processing initially with KNIME, then with the commercial soft-
ware Agisoft Metashape, based on Structure from Motion algorithms.
Subsequently, a model was trained using the YOLO v5s6 algorithm to
identify issues relating to the structure’s joints, made up of metal bolts.
This innovative approach allowed the creation of an automated deep
learning-based method to capture and identify deteriorations, using
bounding boxes. The authors then focused on the implementation of an
application that could allow an immersive and interactive approach to
analyze the different parts of the structure and disseminate related in-
formation. In addition to the practical benets, the application can
contribute to raising awareness about the use of bamboo by dissemi-
nating the opportunities offered by this material among the users who
use it.
Data collection using the drone has proven to be an efcient tech-
nique for obtaining detailed information about the structure. The image
processing time, although considerable (6 hours for the point cloud and
1.2 for the mesh), reects the complexity of the structure and the rich-
ness of the captured results. Future strategies for optimizing substantial
processing time could be the parallel processing to distribute the
workload across multiple CPU cores or GPU units. Parallelizing com-
putations can signicantly reduce processing time. Moreover, the use of
GPUs for accelerating intensive computations can speed up tasks like
point cloud and mesh generation. Another solution could be to upgrade
hardware components like CPU, GPU and memory to improve overall
system performance.
Viewing the point cloud and 3D model illustrates the accuracy ach-
ieved by the process.
Regarding the analysis of junctions, the analysis of temporal frames
allowed detailed comparisons at different instants of time. The imple-
mentation of YOLO v5s6 subsequently enabled the automatic identi-
cation of deteriorated bolts, highlighting the capabilities of the trained
model to accurately identify joint issues and underlining the effective-
ness of deep learning-based approaches in complex structural contexts.
Fig. 6. Comparison of the bolt at time t and time t+1 after applying the edge
detection algorithm present in KNIME.
Fig. 7. Identication of bolts using YOLO v5s6 in the bamboo structure under
examination. The yellow boxes highlight the bolts that present marked deteri-
oration and loosening characteristics.
Fig. 8. (a) App: virtual representation of one stage of the surveyed structure.
(b) App: visualization of 3D model remotely, highlighting the possibility to
interact with it.
V. Barrile and E. Genovese
Advances in Bamboo Science 7 (2024) 100079
8
The use of the app then introduced an element of innovation and
practicality in monitoring the structure. The possibility of consulting
information sheets such as the technical data sheet of the materials or
possibly even a BIM of the structure adds further value to the applica-
tion, facilitating the planning of maintenance interventions.
The innovative monitoring system and Virtual/Augmented/Mixed
Reality (VR/AR/MR) app are fundamental in addressing bamboo’s
durability concerns and structural limitations. The monitoring system
integrates sensors and data analytics to continuously evaluate the con-
dition of bamboo structures. Additionally, remote accessibility enables
engineers and maintenance teams to access real-time data and insights
from anywhere, facilitating proactive decision-making and timely in-
terventions to address structural concerns. Furthermore, predictive
maintenance could be enabled through historical data analysis and trend
prediction, optimizing maintenance schedules to prevent deterioration
and extend the structure’s life. The VR/AR/MR app enhances under-
standing, visualization, and decision-making regarding bamboo struc-
tures through immersive experiences. Users can visualize the internal
and external features of the bamboo structure, including nodes, joints,
and load-bearing components. Simulations of stress and load distribu-
tion help identify potential areas of weakness and optimize structural
design. Collectively, these high-performance aspects of the monitoring
system and VR/AR/MR app contribute to addressing bamboo’s dura-
bility concerns and structural limitations by enabling proactive moni-
toring, predictive maintenance, informed decision-making, and
enhanced understanding of bamboo structures.
5. Conclusion
The research conducted represents a signicant advancement in
structural engineering, particularly in the eld of bamboo structures, by
integrating advanced technologies for detection and analysis. Focused
on a specic case in China, the approach involved meticulous data
collection via drones, processing using specialized software, joint anal-
ysis with advanced algorithms, and the implementation of a virtual,
augmented, and mixed reality app for visualization and interpretation of
results. While the ndings are promising, the study has its limitations,
such as the dependence on weather conditions during drone surveying
and potential challenges in the temporal comparison methodology due
to facility layout changes over time. However, these limitations serve as
avenues for future research developments and improvements. Looking
ahead, optimizing algorithms for enhanced efciency, validating 3D
models through physical measurements, rening detection algorithms
to address varying light conditions, and improving app interactivity for
real-time access to structural information are crucial areas of focus.
These advancements not only contribute to the reliability and suitability
of bamboo structures but also have broader implications for structural
engineering as a whole.
Practically, the research offers tangible benets by enabling
continuous monitoring, analysis, and knowledge expansion of bamboo
as a construction material. By improving advanced technologies, struc-
tural engineers can optimize maintenance strategies, improve structural
integrity, and ensure the safety and sustainability of infrastructure
projects.
Without doubt, this research highlights the transformative potential
of integrating advanced technologies in structural engineering, offering
practical solutions to real-world challenges, and paving the way for
safer, more resilient, and environmentally friendly construction prac-
tices. It underscores the importance of ongoing research and innovation
in advancing the eld and addressing the evolving needs of modern
infrastructure development. This approach highlights the advances in
the eld of bamboo structures, since through continuous monitoring it is
possible to carry out analyzes and above all it allows the expansion of
knowledge of this extremely useful material from an engineering point
of view, further studies and analyzes must be carried out to allow
complete reliability, suitability for any type of structure and real-time
availability.
CRediT authorship contribution statement
Vincenzo Barrile: Writing – review & editing, Supervision, Re-
sources, Project administration, Methodology, Funding acquisition,
Formal analysis, Conceptualization. Emanuela Genovese: Writing –
original draft, Visualization, Validation, Software, Investigation, Data
curation.
Declaration of Competing Interest
The authors declare that they have no known competing nancial
interests or personal relationships that could have appeared to inuence
the work reported in this paper.
Data Availability
Data will be made available on request.
Acknowledgements
The studies presented here comes from the activities envisaged by
FISR/GoForIT/CRUI Foundation Project “Seismic assessment of bridges
and viaducts with Remotely Piloted Aircraft Systems (RPAS) and Arti-
cial Vision” in collaboration with Zhejiang University (China) and the
Department of Engineering Structure of the Sapienza University of
Rome.
References
Alhajahmad, B., Atas
¸, M., Atas
¸, ˙
I., 2023. Utilizing KNIME as a machine learning
classication tool: a practical demonstration with SVM. Random For., Naive Bayes.
-Proc. 1 (1), 223–227.
Arena, F., Collotta, M., Pau, G., Termine, F., 2022. An overview of augmented reality.
Computers 11 (2), 28.
Bala, A., Gupta, S., 2023. Engineered bamboo and bamboo-reinforced concrete elements
as sustainable building materials: A review. Constr. Build. Mater. 394, 132116.
Barrile, V., Candela, G., Bernardo, E., & Fotia, A. (2021). Comparison of Different Digital
Models of a Sustainable Bamboo Structure Using Aerial Photogrammetry. In New
Metropolitan Perspectives: Knowledge Dynamics and Innovation-driven Policies
Towards Urban and Regional Transition Volume 2 (pp. 1302-1309). Springer
International Publishing.
Ben-Zhi, Z., Mao-Yi, F., Jin-Zhong, X., Xiao-Sheng, Y., Zheng-Cai, L., 2005. Ecological
functions of bamboo forest: research and application. J. For. Res. 16, 143–147.
Candela, G., Barrile, V., Demartino, C., & Monti, G. (2019). Image-based 3d
reconstruction of a glubam-steel spatial truss structure using mini-UAV. Modern
Engineered Bamboo Structures; Xiao, Y., Li, Z., Liu, KW, Eds, 217-222.
Chandana, R.K., Ramachandra, A.C., 2022. Real time object detection system with YOLO
and CNN models: A review. arXiv Prepr. arXiv 2208, 00773.
Chang, C.M., Lin, T.K., Chang, C.W., 2018. Applications of neural network models for
structural health monitoring based on derived modal properties. Measurement 129,
457–470.
Cheng, Y., Wan, S., Yao, L., Lin, D., Wu, T., Chen, Y., Lu, C., 2023. Bamboo leaf: A review
of traditional medicinal property, phytochemistry, pharmacology, and purication
technology. J. Ethnopharmacol., 116166
Chaudhry, M.H., Ahmad, A., Gulzar, Q., 2020. Impact of UAV surveying parameters on
mixed urban landuse surface modelling. ISPRS Int. J. Geo-Inf. 9 (11), 656.
Chiang, F.K., Shang, X., Qiao, L., 2022. Augmented reality in vocational training: A
systematic review of research and applications. Comput. Hum. Behav. 129, 107125.
Emamverdian, A., Ding, Y., Ranaei, F., Ahmad, Z., 2020. Application of bamboo plants in
nine aspects. Sci. World J. 2020.
Escamilla, E.Z., Habert, G., Santos, H.A., Fern´
andez, J.S.E., Trujillo, D., 2018. Industrial
or traditional bamboo construction? Comparative life cycle assessment (LCA) of
bamboo-based buildings. Sustainability 10 (9), 3096.
Fahim, M., Haris, M., Khan, W., Zaman, S., 2022. Bamboo as a construction material:
Prospects and challenges. Adv. Sci. Technol. Res. J. 16 (3).
Fotia, A., Barrile, V., 2023. Viaduct and Bridge Structural Analysis and Inspection
through an App for Immersive Remote Learning. Electronics 12 (5), 1220.
Fotia, A., Pucinotti, R., Barrile, V., 2022. Detection of Steel Structures Degradation
through a UAVs and Articial Intelligence Automated System. WSEAS Trans. Circuits
Syst. 21, 231–237.
Ghavami, K., Moreira, L.E., 1996. Development of a new joint for bamboo space
structures. WIT Trans. Built Environ. 24.
Godone, D., Allasia, P., Borrelli, L., Gull`
a, G., 2020. UAV and structure from motion
approach to monitor the maierato landslide evolution. Remote Sens. 12 (6), 1039.
V. Barrile and E. Genovese
Advances in Bamboo Science 7 (2024) 100079
9
Greco. (2017). S. Il Bambù italiano: propriet`
a meccaniche e connessioni assiali per
prospettive sostenibili. Master’s Degree’s Thesis in Construction Science,
Construction Engineering and Architecture.
Gu, H., Yang, M., Gu, C.S., Huang, X.F., 2021. A factor mining model with optimized
random forest for concrete dam deformation monitoring. Water Sci. Eng. 14 (4),
330–336.
Gupta, Y. (2023). Structural Health Monitoring of the Marine Structure Using Guided
Wave and UAV-based testing.
Han, S., Chen, F., Yu, Y., Chen, L., Wang, G., 2023. Bamboo-inspired strong, tough and
stable composites derived from renewable bamboo. Ind. Crops Prod. 194, 116292.
Hasan, S.M., Lee, K., Moon, D., Kwon, S., Jinwoo, S., Lee, S., 2022. Augmented reality
and digital twin system for interaction with construction machinery. J. Asian Archit.
Build. Eng. 21 (2), 564–574.
Hegde, A., Sitharam, T.G., 2015. Use of bamboo in soft-ground engineering and its
performance comparison with geosynthetics: experimental studies. J. Mater. Civ.
Eng. 27 (9), 04014256.
Heichel, J., Mitra, R., Jafari, F., Das, A., Dorafshan, S., Kaabouch, N., 2023. June). A
System for Real-Time Display and Interactive Training of Predictive Structural
Defect Models Deployed on UAV. In 2023. International Conference on Unmanned
Aircraft Systems (ICUAS). IEEE, pp. 1221–1225.
Ivi´
c, S., Crnkovi´
c, B., Grbˇ
ci´
c, L., Matlekovi´
c, L., 2023. Multi-UAV trajectory planning for
3D visual inspection of complex structures. Autom. Constr. 147, 104709.
Janssen, J.J., Boughton, G., Adkoli, N.S., Ranjan, M.P., Sastry, C.B., Ganapathy, P.M.,
Ravindran, K. (1981). Bamboo as an engineering material. The IDRC Bamboo and
Rattan Research Network, Eindhoven University.
Jiang, P., Ergu, D., Liu, F., Cai, Y., Ma, B., 2022. A Review of Yolo algorithm
developments. Procedia Comput. Sci. 199, 1066–1073.
Jitendra, M.S., Srinivas, A.S., Surendra, T., Rao, R.V., Chowdary, P.R. (2021, October). A
study on game development using unity engine. In AIP Conference Proceedings (Vol.
2375, No. 1). AIP Publishing.
Juliani, A., Berges, V.P., Teng, E., Cohen, A., Harper, J., Elion, C., Lange, D. (2018).
Unity: A general platform for intelligent agents. arXiv preprint arXiv:1809.02627.
Karthika, S., Praveena, P., GokilaMani, M., 2017. Hololens. Int. J. Comput. Sci. Mob.
Comput. 6 (2), 41–50.
Kleinsmann, J., Verbesselt, J., Kooistra, L., 2023. Monitoring Individual Tree Phenology
in a Multi-Species Forest Using High Resolution UAV Images. Remote Sens. 15 (14),
3599.
Kohli, V., Tripathi, U., Chamola, V., Rout, B.K., Kanhere, S.S., 2022. A review on Virtual
Reality and Augmented Reality use-cases of Brain Computer Interface based
applications for smart cities. Microprocess. Microsyst. 88, 104392.
Kumar, R., Ganguly, A., Purohit, R. (2023). Properties and applications of bamboo and
bamboo bre composites. Materials Today: Proceedings.
Lan, W., Dang, J., Wang, Y., Wang, S., 2018. August). Pedestrian detection based on
YOLO network model. In 2018. IEEE International Conference on Mechatronics and
Automation (ICMA). IEEE, pp. 1547–1551.
Liu, C., Tao, Y., Liang, J., Li, K., Chen, Y., 2018. December). Object detection based on
YOLO network. In 2018. IEEE 4th Information Technology and Mechatronics
Engineering Conference (ITOEC). IEEE, pp. 799–803.
Lou, Z., Wang, Q., Sun, W., Zhao, Y., Wang, X., Liu, X., Li, Y., 2021. Bamboo attening
technique: A literature and patent review. Eur. J. Wood Wood Prod. 79, 1035–1048.
Majumder, M., Wilmot, C., 2023. Automated vehicle counting from pre-recorded video
using you only look once (YOLO) object detection model. J. Imaging 9 (7), 131.
Martins, N.C., Marques, B., Alves, J., Araújo, T., Dias, P., Santos, B.S., 2022. Augmented
reality situated visualization in decision-making. Multimed. Tools Appl. 81 (11),
14749–14772.
Moran, R., García, J.J., 2019. Bamboo joints with steel clamps capable of transmitting
moment. Constr. Build. Mater. 216, 249–260.
Moiseienko, N.V., Moiseienko, M.V., Kuznetsov, V.S., Rostalny, B.A., Kiv, A.E., 2023.
Teaching computer game development with Unity engine: a case study. CEUR
Workshop Proc. 237–251.
Park, S., Bokijonov, S., Choi, Y., 2021. Review of microsoft hololens applications over the
past ve years. Appl. Sci. 11 (16), 7259.
PyTorch, accessed on 26 April 2023. https://pytorch.org/hub/ultralytics_yolov5/.
Seixas, M., Moreira, L.E., Stoffel, P., Bina, J., Ripper, J.L.M., Ferreira, J.L., Ghavami, K.,
2021. Analysis of a self-supporting bamboo structure with exible joints. Int. J.
Space Struct. 36 (2), 137–151.
Sylvayanti, S.P., Nugroho, N., Bahtiar, E.T., 2023. Bamboo Scrimber’s Physical and
Mechanical Properties in Comparison to Four Structural Timber Species. Forests 14
(1), 146.
Sharma, B., Gatoo, A., Bock, M., Mulligan, H., Ramage, M., 2015. Engineered bamboo:
state of the art. Proc. Inst. Civ. Eng. -Constr. Mater. 168 (2), 57–67.
Shu, B., Xiao, Z., Hong, L., Zhang, S., Li, C., Fu, N., Lu, X., 2020. Review on the
application of bamboo-based materials in construction engineering. J. Renew. Mater.
8 (10), 1215–1242.
Terven, J., Cordova-Esparza, D., 2023. A comprehensive review of YOLO: From YOLOv1
to YOLOv8 and beyond. arXiv Prepr. arXiv 2304, 00501.
Wang, Y.Y., Wang, X.Q., Li, Y.Q., Huang, P., Yang, B., Hu, N., Fu, S.Y., 2020. High-
performance bamboo steel derived from natural bamboo. ACS Appl. Mater.
Interfaces 13 (1), 1431–1440.
Wang, Z., Han, K., Tiwari, P. (2021, July). Digital twin simulation of connected and
automated vehicles with the unity game engine. In 2021 IEEE 1st International
Conference on Digital Twins and Parallel Intelligence (DTPI) (pp. 1-4). IEEE.
Widyowijatnoko, A., Harries, K.A. (2020). Joints in bamboo construction. In
Nonconventional and Vernacular Construction Materials (pp. 561-596). Woodhead
Publishing.
Xu, X., Xu, P., Zhu, J., Li, H., Xiong, Z., 2022. Bamboo construction materials: carbon
storage and potential to reduce associated CO2 emissions. Sci. Total Environ. 814,
152697.
Yadav, M., Mathur, A., 2021. Bamboo as a sustainable material in the construction
industry: an overview. Mater. Today.: Proc. 43, 2872–2876.
Zhang, S., Jin, H., Feng, X., Pan, W., Chen, Y., Jin, B., Zhang, H., 2023. Inuence of
NaOH concentration and hydrothermal treatment on the morphology and properties
of wood bers in wood plastic composites. Compos. Part B: Eng., 119310
Zhou, C., Chase, J.G., Rodgers, G.W., 2021. Support vector machines for automated
modelling of nonlinear structures using health monitoring results. Mech. Syst. Signal
Process. 149, 107201.
Zhong, C., Zhang, Y., Li, Y., Fan, H., 2022. A review of structural health monitoring for
civil engineering based on articial intelligence techniques. J. Civ. Struct. Health
Monit. 12 (1), 163–190.
Zhu, J., Chu, H., Li, Y., Wang, L., Yan, S., 2023. A comparative study on the carbon
footprint and energy consumption of bamboo and concrete formwork in the
construction industry. J. Clean. Prod., 125532
V. Barrile and E. Genovese
