Research Paper

Advances in Structural Engineering

2025, Vol. 28(1) 122–144
A scientometric analysis of drone-based © The Author(s) 2024

structural health monitoring and new Article reuse guidelines:

sagepub.com/journals-permissions
technologies DOI: 10.1177/13694332241255734
journals.sagepub.com/home/ase

Tahreer M Fayyad1 , Su Taylor1, Kun Feng2  and Felix Kin Peng Hui3

Abstract
Critical global challenges, such as climate change and the insufficient availability of resources, mean that it is a pivotal time to
make cities more intelligent, efficient, and sustainable in a drive towards a net-zero carbon future. This requires intelligent,
interactive, and responsive structural health monitoring (SHM) to assure the longevity and safety of ageing infrastructure.
Drones have the potential to revolutionise SHM. Drone-based SHM (as a potential fly-by technique) involves equipping
drones with various sensors, or using inbuilt sensors, to capture data and images of structures from different angles and
perspectives. The data is then processed and analysed to facilitate accurate assessment of the structure’s health and early
diagnosis of damage. Although the use of fly-by is relatively new, the speedy advances in various technologies that could be
integrated with it, such as computer vision with artificial intelligence, deep learning, and links to digital twins, put these
systems on the verge of a potential breakthrough. This paper provides an overview of fly-by SHM technique using both
scientometric and qualitative systematic literature review processes, in order to provide a distinct understanding of the
state of the art of research. As an original contribution, our research identified four main clusters of research within the
field of fly-by SHM: (1) the application of UAV-enabled vision-based monitoring; (2) the integration of drones, advanced
sensor technologies, and artificial intelligence; (3) drone-based SHM integrating modal analysis, energy harvesting, and deep
learning; and (4) automation and robotics in drone-based SHM. The paper highlights the integration of new technologies
such as artificial intelligence, machine learning, and sensors with the fly-by technique for SHM, identifies the gaps in current
fly-by SHM research, and suggests new directions for research.

Keywords
unmanned aerial vehicles, drones, structural health monitoring, intelligent infrastructure, fly-by, sensors, imaging, artificial
intelligence, machine learning, scientometric analysis

Introduction system, which acts as a highly advanced information-

gathering system, sensing the environment and provid-
Structural health monitoring (SHM) is a growing field in ing information about hidden problems through symptoms,
civil engineering and can play a pivotal role in maintaining thus allowing for appropriate action to keep the body safe.
and ensuring the longevity and safety of critical civil en- This can be compared to structures where sensors can be
gineering infrastructure. Damage in civil engineering in-
frastructure can be an accumulative process which affects
1
the structure until it reaches a point of losing the ability to Intelligent and Sustainable Infrastructure Group, School of Natural and
function safely. For example, the corrosion of steel re- Built Environment, Queen’s University Belfast, Belfast, UK
2
School of Mechanical and Aerospace Engineering, Queen’s University
inforcing bars in concrete accelerates after initiation.
Belfast, Belfast, UK
Hence, the significance of early diagnosis in structures is 3
Engineering Management Group, Department of Infrastructure
similar to that in healthcare, where early and accurate Engineering, University of Melbourne, Melbourne, VIC, Australia
diagnosis of illness can be significant for timely inter-
vention, enabling tailored treatment that can improve a Corresponding author:
Tahreer M Fayyad, Intelligent and Sustainable Infrastructure Group,
patient’s prognosis and expedite recovery. Human bodies School of Natural and Built Environment, Queen’s University Belfast,
are magnificent structures that can feel and detect changes Belfast BT7 1NN, UK.
and react accordingly. This is due to the human nervous Email: [email protected]
Fayyad et al. 123

provided to monitor changes such as pressure, temperature, occurrence of wildfires, often associated with heatwaves,
and vibration. These sensors would provide information poses direct threats to structures (Dennison et al., 2014). As
about hidden deficiencies, give warnings, save lives, assure the threat of global warming intensifies, early detection and
safety, and enable informed strategic budget allocation for ongoing monitoring offer a chance to extend the life and
repair and maintenance (Figure 1). In the last two decades, safety of critical infrastructures.
there have been more than a hundred major bridge col- Structural health monitoring (SHM) is the technical
lapses worldwide (Zhang et al., 2022). In 2018, the method of non-destructively assessing various aspects
Morandi Bridge in Italy collapsed, causing 43 deaths and related to the condition of a structure during its operation.
significant economic damage (Calvi et al., 2019). Hence, This encompasses the gathering and analysis of informa-
monitoring the network of civil engineering infrastructure tion pertinent to the health and integrity of structures. It
over time and extracting valuable information via SHM can involves assessing safety ratings, damage, and employing
act as an early warning and prevent reaching the critical whole-life prediction modelling. SHM is integral to en-
point of failure during the service life of structures (Bao suring the longevity and safety of engineering systems by
et al., 2019; Li et al., 2006; Lydon et al., 2019b). The fi- providing critical insights into their current condition and
nancial cost of repairing faults as they approach the critical projecting their future performance (Douglas E. Adams,
phase is enormous, but if damage is prevented at an early 2007). Sensing is an essential aspect of SHM. The specific
stage, informed decisions can be made before the structures goals of an SHM initiative, along with factors such as the
are damaged beyond repair (Zonta et al., 2014). Informed nature and scale of the civil structures involved, their
decision-making can also assist in the allocation of geographic positions, accessibility, prevailing weather
stretched budgets for prioritising the maintenance of the conditions, and economic considerations, dictate the se-
most critical assets. Global warming poses new challenges, lection of sensing technologies. Both contact and non-
with more frequent and extreme events, increasing the risks contact sensors, supported by wired and wireless net-
to our infrastructure. Heatwaves can induce thermal ex- works, including Internet of Things (IoT) platforms, are
pansion and raise the incidence of material cracking that primarily utilised. These systems are designed to capture a
threatens the structural integrity (Hossain et al., 2020; Zhou wide range of structural behaviours as well as significant
and Yi, 2013). Moreover, the materials used in constructing environmental and operational information, facilitating the
these infrastructures can deteriorate faster under flood transmission of this data to both local storage solutions and
events, increasing the potential for scour in bridges that cloud-based servers for analysis and decision-making
could lead to serious failure, such as what happened with (Sarmadi et al., 2023). Contact sensors, including accel-
the Morandi Bridge (Calvi et al., 2019). The increasing erometers, strain gauges, piezoelectric transducers, fibre

Figure 1. SHM to assess structure health (analogous to medical testing of humans).

124 Advances in Structural Engineering 28(1)

optic sensors, linear variable differential transformers, thermal, and LiDAR sensors, providing a comprehensive
thermocouples, and anemometers, among others, are di- perspective on structural health (Poorghasem and Bao, 2022;
rectly integrated into civil structures. These sensors are Sony et al., 2019; Wang and Ueda, 2023). Therefore, the use
deployed to capture key structural reactions, such as ac- of drones can address the demands of traditional SHM
celeration, strain, and displacement, or to monitor envi- techniques, making them safer, more efficient, and more
ronmental conditions, including temperature and wind integrated, leading to more effective structural inspections.
speed and direction. Conversely, non-contact sensors The market for drones is expected to double in the next 3 years
represent a more recent addition to the field of SHM, (Mohsan et al., 2022). This growth is driven by the increasing
functioning without direct attachment to the structures demand for drones in different applications (Chen et al.,
themselves. These sensors, typically operating from a 2016). SHM in various industries such as civil engineer-
distance, are employed to capture optical images and ing, oil and gas, and wind energy is one of these promising
videos. They utilise a range of technologies, from com- applications. Drone-based SHM (UAV-based SHM) is one of
mercial digital and high-speed cameras to video cameras, the most recent techniques for SHM, as it can be used in a
as well as optical and synthetic aperture radar imagery variety of environments, including urban, rural, and remote
obtained through satellite sensors. This technology in- areas. They can now be used for more than just inspecting
cludes smartphone sensing technology. large and complex structures, such as bridges, wind turbines,
More recently, there have been advances in technologies and high-rise buildings that may be difficult or dangerous to
that can change the way we live by moving towards ‘smart’ access using traditional inspection methods. The use of vision
cities, where services and networks are more efficient sensors enables the detection of displacements, strains, and
through the use of digital and telecommunication tech- crack openings. Drones passing over or flying around a
nologies (Khan et al., 2016; Li et al., 2006; Ozer and Feng, structure can capture data and images, which are used in
2019; Xu et al., 2016; Zhu and Law, 2015). These tech- conjunction with other vision-based techniques, such as
nological advances coexist with the need to adapt to computer vision feature tracking, photogrammetry, LiDAR,
sustainable solutions that ensure efficient consumption of or infrared thermal (IRT) imaging- where IRT can detect
our natural resources and make cities more intelligent and hidden defects (Pedram et al., 2024) - to collect compre-
efficient while driving towards our net-zero carbon future. hensive and accurate data about the changing health of a
This requires intelligent infrastructure through interactive structure (Figure 2) (Efstathios et al., 2021; E Polydorou et al.,
and responsive SHM, which is on the verge of a break- 2021; Poorghasem and Bao, 2022; Rathinam et al., 2008; Tse
through when integrated with digital transformation in the et al., 2023; Wang and Ueda, 2023). Accelerometers, in-
monitoring, processing, and analysis of information in real cluding built-in systems, can also be used in drones for SHM
time (Noel et al., 2017; Tokognon et al., 2017). Effective in various ways (Feng et al., 2019). By measuring the vi-
SHM requires comprehensive and precise data collection, brations of a structure using accelerometers on a UAV, it is
often from hard-to-reach or hazardous areas. It demands possible to detect changes in the modal frequency due to
regular and consistent monitoring to capture any degra- stiffness alterations resulting from damage (Feng et al., 2019;
dation progress or process due to the effect of changing Tse et al., 2023). The change in the structure’s modal fre-
loading or environmental conditions on the structure. quency can be used to quantify damage, such as cracks or
Moreover, real-time or near real-time data processing is deformation, and to inform detailed 2D or 3D models of the
desirable for rapid assessment and decision-making. This is structure for analysis and simulation. UAVs have the potential
where new technologies play a vital role in future SHM (Ko to revolutionise SHM of buildings, bridges, wind turbines,
and Ni, 2005; Lydon et al., 2012; Pedram et al., 2021). and other critical infrastructures. This technique of collecting
Drone or Unmanned Aerial Vehicle (UAV) technology has information about the health of structures from the sensor data
the potential to meet the requirements for intelligent infra- collected by drones is referred to as fly-by in this review. It has
structure by enhancing the effectiveness and integration of the potential to transform how the health of structures is
SHM techniques. Equipped with a variety of sensors, drones monitored, identifying defects, and predicting potential fail-
can access hard-to-reach areas and reduce human exposure to ures in real-time. Inspection of our bridge infrastructure is
potentially dangerous situations. They enable structured and usually periodic and involves examinations of structures to
repeatable inspections, which are crucial for maintaining detect visible signs of damage or degradation. It relies on the
temporal consistency in data collection throughout the life- expertise of the inspector to assess the state of the structure,
span of a monitoring project. A built-in accelerometer can also using tools and methods that provide point-in-time evalua-
serve as a sensor for modal analysis if the drone is ‘parked’ on tions and are often subjective and prone to human error
a bridge. The data collected by drones can be processed, (Pedram et al., 2024). However, SHM is a continuous or semi-
enabling real-time or near real-time structural assessments continuous process that employs data collection and data
(Kressel et al., 2015; Liang et al., 2023). Drones can facilitate analysis techniques to monitor the condition and performance
the integration of various sensor modalities, including visual, of a structure, potentially in real-time, for ongoing assessment.
Fayyad et al. 125

Figure 2. The integration of drones with different sensing system for SHM.

Drones have been used successfully in the field for inspection 2015). This approach identifies potential connections be-
tasks, however, with recent technologies, it is now being tween literature concepts that might otherwise go unno-
integrated into more advanced SHM that provides timely ticed in manual review studies, enabling a more
insights, coupling traditional inspection tasks with cutting- comprehensive overview of the field, which will yield new
edge SHM methodologies. insights into the developing field. For this scientometric
The rapidly increasing demand for drones in SHM has review, the specific objectives were: (1) analysing the
attracted the interest of the research community. Although stature of drone-based SHM, namely, “fly-by” SHM; (2)
there is no review on drone-based SHM, recent reviews in analysing the trends and nature of academic publications in
the broader field of SHM some insights into the potential the area of drone-based SHM; and (3) science mapping of
use of drones (Ellenberg et al., 2015; Liang et al., 2023; publication outlets and keywords and defining research
Nooralishahi et al., 2021; Poorghasem and Bao, 2022; clusters. For the further critical systematic literature review,
Reagan et al., 2018; Sony et al., 2019; Tian et al., 2022; the objective was to explore in more depth the literature
Wang and Ueda, 2023). from the scientometric analysis, focusing on the method-
While traditional literature reviews offer significant ologies, findings, and implications in the context of drone-
insights, their qualitative and manual nature can introduce a based SHM. This further in-depth analysis is presented
degree of subjectivity. To the authors’ knowledge, there is within the scientometric analysis results to ensure a
an evident research gap where no reviews have been comprehensive understanding of the latest research,
conducted on drone-based SHM, nor have any studies highlighting the trends and the gaps, thus leading to a more
sought to establish the intricate connections or operational comprehensive overview than either method alone can
dynamics among clusters of researchers, journals, and provide. It should be noted that the main aim of this paper is
countries engaged in drone-based SHM. Furthermore, the to present a scientometric quantitative review, where the
domain lacks a comprehensive analysis that encompasses reader is directed to the references being cited to gain more
keyword co-occurrence, evolution, and cluster examina- technical depth behind each cluster in the field.
tion. Recognising these limitations, this paper aims to
deliver an unbiased review of drone-based SHM. To
achieve this, a combined approach that includes integrating
Research methodology
scientometric analysis with a further critical review is A mixed review approach, scientometric and qualitative
presented. Scientometric methods enable systematic systematic literature review, was employed to provide a
literature-related findings (Mingers and Leydesdorff, distinct understanding of the state of the art of research of
126 Advances in Structural Engineering 28(1)

drone-based SHM. The technique of scientometric anal- Meanwhile, the objective of the critical review is to un-
ysis, commonly employed for inspecting the advancement derstand and present research themes and corresponding
of research, offers a lens to examine the scholarly pro- challenges using the scientometric findings as a basis.
ductivity of a country, individuals, academic departments, Figure 3 shows the methodology diagram.
and academic journals (Mingers and Leydesdorff, 2015) The Scopus database was selected as the literature re-
Scientometric analysis emerges from the intersection of source because it offers a more comprehensive coverage of
information science and bibliometric metrics which focus structures-related research compared to other databases
on a quantitative study and assessment of scientific (Mongeon and Paul-Hus, 2016). The scientometric anal-
research. Where publications can be measured and ana- ysis was performed in two phases; first, it was conducted on
lysed, the scientometric approach seeks to uncover pat- the SHM field in general to determine the stature of UAV-
terns, trends, and networks in scientific literature. The based SHM within the field, and then it focused on the
process typically begins with the extraction of biblio- specific topic of drone-based SHM. The analysis in this
graphic data from academic databases, followed by rig- paper was performed using Scopus analyser, Excel and
orous data cleansing to ensure accuracy and consistency. VOSviewer software tool for science mapping (Ding and
Subsequent stages involve computational and statistical Yang, 2022). VOSviewer was chosen due to its features and
analyses, using tools and techniques ranging from co- capabilities that align well with the research’s objectives. It
citation analysis to network visualisation, offering in- is a tool designed to generate, visualise, and investigate
sights into the structure and dynamics of scientific disci- networks in bibliometrics. Systematic literature review was
plines (Ivancheva, 2008; Mingers and Leydesdorff, 2015). done using the Preferred Reporting Items for Systematic
By charting the flow of research topics and mapping the Review and Meta-Analyses (PRISMA) to enhance the
interconnectivity among researchers and keywords, sci- reporting quality of review.
entometric analysis provides an overall quantitative view of
the evolution and current state of a particular scientific field
(Ding and Yang, 2022; Van Eck and Waltman, 2017). It has
Scientometric quantitative analysis
been widely used to track the chronological progression of A basic search was conducted on Scopus for all SHM
research in different fields including construction. This publications. After several rounds of refinement, the
includes areas such as sustainable development (Olawumi field was limited to Engineering to exclude publications
and Chan, 2018), the application of artificial intelligence related to medical or other fields. Additionally, the
(Darko et al., 2020), computer vision applications language was restricted to English to facilitate tracking
(Martinez et al., 2019), and building information modelling keywords without affecting link strength calculations
(He et al., 2017). This specific approach is employed in this during the analysis. The search query was: TITLE-ABS-
paper in the context of drone-based SHM to map out the KEY (structural AND health AND monitoring) AND
area of scientific knowledge systematically and objectively. (LIMIT-TO ( SUBJAREA, “ENGI”)) AND (LIMIT-TO

Figure 3. Research methodology diagram.

Fayyad et al. 127

( LANGUAGE, “English”)) AND (EXCLUDE (EX- (Van Eck and Waltman, 2014) allocates nodes in a network
ACTKEYWORD, “Biological Systems”)). The search into clusters, with each cluster representing a collection of
resulted in finding 30,466 documents and this includes closely interconnected nodes. Each node in the network is
conference papers, reviews, book chapters and articles. exclusively associated with a single cluster. The quantity of
The published articles provide a reflection on the de- clusters is established through a resolution parameter,
velopment of a field and how mature the research is. where a higher value leads to an increased count of clusters.
These initial results were screened for articles When visualisation a bibliometric network, VOSviewer
(13,584 articles) which are presented in Figure 4. The employs distinct colours to signify the cluster affiliation of
results show that most of the publications (96.955%) each node. The method employed for clustering within
were published within the span of the last two decades, VOSviewer is the one that introduced by Waltman and Van
from 2003 to 2023. Figure 5 shows the distribution of Eck (Van Eck and Waltman, 2014, 2017). The analysis of
article publications by country/territory across all as- keyword co-occurrence was conducted leading to the
pects of SHM, with China and the United States with the identification of 12 distinct clusters (Figure 6). Defining the
highest volume followed by United Kingdom and Italy. specific clusters within the entire field of Structural Health
To better understand the more recent evolution of UAV Monitoring (SHM) requires deep analysis and substantial
integration within SHM, an analysis of the last decade was argumentation that is beyond the scope of the current paper.
conducted (Mohsan et al., 2022). This data was exported to However, the primary aim of Figure 6 and the associated
VOSviewer and the research themes were determined analysis was to identify the position of UAVs within the
using the keywords co-occurrence analysis. VOSviewer broader SHM field.

Figure 4. Publications in structural health monitoring over the years.

Figure 5. Publications in structural health monitoring by country/territory.

128 Advances in Structural Engineering 28(1)

Figure 6. Keyword mapping for SHM, 2013-2023.

UAVs/drones were found to be among one of the largest (PRISMA). PRISMA represents a minimum evidence-
clusters in SHM (the red cluster in Figure 6), and this based set of items designed for the reporting in system-
includes all the words that refer to UAVs, such as “drone”, atic reviews and meta-analyses. PRISMA’s objective is to
“unmanned aerial vehicle/s”, and “UAV”. A zoom into the assist authors in enhancing the reporting quality of sys-
red cluster is shown in Figure 7. The centrality of UAVs in tematic reviews and meta-analyses (Pedram et al., 2024).
the one of the biggest SHM clusters reinforces their integral Figure 9 below shows the entire process, the number ar-
role in shaping the future of infrastructure maintenance and ticles that are screened and excluded for a final set of items
safety. to be analysed.
Further analysis is presented in the subsequent sections For UAV-based SHM, an initial search was carried out,
regarding the relationships between UAV and other inter- which was later fine-tuned using prevalent keywords. The
related keywords within the field, the findings also show initial search string included all the keywords that point to
that UAVs began to play a pivotal role in the field of SHM UAVs as follows: TITLE-ABS-KEY (structural AND
over the last decade. This observation is further supported health AND monitoring AND ((unmanned AND aerial
by investigating the overlay visualization (Figure 8), which AND vehicle OR uav OR drone) AND (LIMIT-TO
illustrates the temporal frequency of all the words/terms (LANGUAGE, “English”)) AND (LIMIT-TO ( SUB-
over time. It is clear that the words related to UAV are JAREA, “ENGI”)). This search resulted in 417 publica-
coloured to show increased prevalence from 2020 onwards tions. The type of publication is shown in Table 1. These
(yellow and green). Their prominence in recent literature categories are taken as they are in Scopus that cannot be
not only highlights the technological advancements in the changed but can be filtered and investigated more.
area but also underscores the growing recognition of their The articles were subjected to further reading by the first
utility. author to check that it matches the objective of the review.
Systematic literature review is done using the Preferred While many of the papers concentrated on the research or
Reporting Items for Systematic Review and Meta-Analyses development of the UAVs themselves and the technologies
Fayyad et al. 129

Figure 7. Drone/UAV in keyword mapping SHM 2013-2023.

Figure 8. Overly visualisation for SHM 2013–2023.

130 Advances in Structural Engineering 28(1)

Table 2. Types of Publications in the Field With UAVs and SHM.

Type of publications No.

Conference paper 104

Article 57
Conference review 9
Review 9
Book chapter 3

and for quality control in construction where UAVs gather

images at different phases of the construction process to
supervise project progression. This refined search resulted
in 182 articles exported to Excel for analysis and to
VOSviewer for science mapping. The refined output of
publication is shown in Table 2. The nine papers that are
classified as reviews were subjected to further investigation
and reading, four of these nine papers were excluded for not
being related specifically to UAV-SHM, however, they
mention SHM and damage detection as a possible appli-
cations with drones. Further to the additional manual check
using more traditional techniques (such as Google scholar),
two more papers were added which are very recent papers.
The content analysis of the review papers (as will be ex-
Figure 9. PRISMA Flow diagram.
plained more in section 6) are reviews of areas that in-
tersects with drone-based SHM. They overlap in different
Table 1. Types of Publications in the Field of Drones -SHM. ways with the fly-by technique and they will be presented
in section 6 to provide more resources for the reader.
Type of publications No.

Conference paper 217 Publication trends

Journal paper/Article 128
Conference review 43 Figure 10 shows the annual variation in the number of
Review 17 publications - both in journals and conference proceedings.
Book chapter 6 A general rise in the publications describing research using
drones for SHM has been observed since 2010, with two
significant surges in 2017-2018 and 2021-2022 which
of developing their manufacturing, often mentioning SHM coincided with the rapid advancements in UAV technology
as a potential application, the core research wasn’t SHM. and big data techniques. However, a noticeable decline in
Even though these technological advances form the bed- publications occurred in 2020, likely attributed to the
rock for UAV-integrated SHM, taking it to new levels, they COVID-19 pandemic’s impact. Also, investigating the
diverge from the central aim of this review, which em- total number of publications in all the fields related to
phasizes the utilisation of UAV based SHM. Improvements UAVs in the Scopus database displays an exceptional
in terms/keywords were made through different rounds of plateau or flat trend in 2020 (Figure 11). Therefore, the
refinement, and the search query was progressively refined results in Figure 10 match the overall trend. It should be
across several rounds, leading to improved investigation mentioned that for the year 2023, the study takes into
results. The resulting refined search string is as follows: account publications from the first 5 months only. If a linear
TITLE-ABS-KEY (“UAV” OR “drone” OR “unmanned regression is drawn, the rising trend continues for 2023,
vehicle” AND “structural health monitoring” AND “in- predicting over 40 publications on fly-by SHM.
spection” OR “assessment” OR “damage detection” OR
“flyby” OR “fly-by”) AND (LIMIT-TO ( SUBJAREA,
“ENGI”)). The search string used here specifically targets
Research outlets
the field of SHM, thereby excluding the various other The 182 publications, identified from the refines search,
applications of UAVs within the broader construction or include 66 journal papers and the majority of them are
building domains, such as human inspection from images published in six academic journals; Structural Health
Fayyad et al. 131

Monitoring, Journal of the International Measurement presenting research on in fly-by SHM, only 9 fulfil the
Confederation, Automation in Construction, Structural criteria. The collaboration network is shown in Figure 13.
Control and Health Monitoring, Sensors and Applied The United States of America (USA) is the most productive
Sciences (Switzerland) as listed in Table 3. country with 70 publications and 1033 citations. This might
be because of the early adoption of large research projects
in UAVs for inspection that lead to start developing it more
Mapping of countries
for SHM applications.
Recognising the most productive countries in a specific One of the earliest practical applications of UAV-based
research domain can bolster network collaboration. This inspection was in the USA in 2015 (Zink and Lovelace,
insight is vital for guiding research funding and nurturing 2015) within a research project investigating inspection of
enhanced international partnerships. Figure 12 illustrates bridges sponsored by the Minnesota Department of
the publication distribution by country. There is evident Transportation. The project assessed four different bridges
interest and significant research effort in the United States across Minnesota, while evaluating the effectiveness of
specifically towards the fly-by SHM technique. Many sensor technology on UAVs, safety, and practicality “ac-
countries show a noticeable equal interest in the field such curacy” compared to other inspection methods. The UAVs
as South Korea, China, Italy and Spain (Figure 12). For were equipped with various imaging technologies, in-
Bibliographic coupling, the search criteria adopted in cluding still RGB image, video, and IRT sensors, to fa-
VOSviewer include analysis type (co-authorship), and cilitate inspection. Some additional data was also collected
analysis unit (country), while the minimum number of during the inspections, site maps, and 3D models of the
documents of a country and the minimum number of ci- bridge elements. The research project indicated that UAVs
tations of a country were set to 7 and 10, respectively. can safely be used during bridge inspections, proving
Based on this research criteria, out of 34 countries especially useful for larger bridges. Another project tar-
geted the Placer River Trail bridge in Alaska in 2017
(Khaloo et al., 2018a). The findings demonstrated the
ability of UAV vision based techniques to produce superior
3D models to support decision-making. While SHM needs
more accuracy to be achieved for practical and effective
SHM, this inspection project was the start for more ad-
vanced usage of drones within the field.
Additionally, Figure 13 identifies three country clusters
based on their citation frequency. For example, the USA,
Australia, and Italy form one cluster, indicated in red. The
other two country clusters are coloured blue and green.
Countries positioned closely in this figure, like Spain and
South Korea, often cite each other. The thickness of the
connecting lines signifies higher citation frequency. Lines
Figure 10. Annual variation of UAV-based SHM publications. between countries appear closely packed, likely because

Figure 11. Publications on Scopus on UAVs in all fields.

132 Advances in Structural Engineering 28(1)

Table 3. Academic Journals Production.

Journal Articles Citations Total link strength

Structural health monitoring 6 188 8

Measurement: Journal of the international measurement confederation 5 141 5
Automation in construction 5 54 8
Structural control and health monitoring 4 273 6
Sensors 4 7 7

Figure 12. Publication distributions of drone-based SHM between countries.

Figure 13. Collaboration network in UAV-based SHM.

Fayyad et al. 133

the field is relatively new and has a limited number of clusters were accurately represented, the occurrence
publications. While USA has the most citations, the benchmark was established to be greater than one and
VOSviewer overlay visualisation shows the China and some refinements were done on the words such as typos.
Australia have the most recent ones. The cluster resolution was reduced to 0.4 to develop four
distinct clusters as shown in Figure 14. The clusters are
developed based on the degree of co-occurrence and how
Keywords co-occurrence analysis
frequently they appear together. The authors went looked
The keywords chosen by the authors for their publica- at these clusters and thought of how these clusters can
tions and in the abstract frequently represent the main reflects the different research themes in the field of study.
themes of a scholarly paper. These keywords are typi- These clusters are discussed in the following section.
cally used for indexing to databases for enhancing search
efficiency. Consequently, a network of these keywords
can potentially depict a specific field of knowledge and Research fields of drone-based SHM
provide insight into vital research topics. A keyword co- (fly-by)
occurrence analysis was carried out to build and illus-
trate the knowledge relationship between drone and Clusters derived from keyword co-occurrence analysis
SHM. The dimension of an element corresponds to its highlight close associations between different keyword
significance, with the label size is proportional to the groups and their citation frequencies. This provides
number of publications where the keywords appear insight into key research topics in this area. These
(Figure 14). The proximity between two keywords in- findings were further investigated within the specific
dicates the degree of their co-occurrence link, a measure research areas:
of how frequently items (in this case, keywords) appear
together in a document. The robustness of this co- Cluster (1): Vision-based structural health
occurrence connection is decided by the number of
publications where the two terms appear jointly. The
monitoring using unmanned aerial Vehicles
keywords were subsequently grouped into clusters, each A UAV is fundamentally designed around a vision
representative of different knowledge domains using system, which includes a camera (e.g., infrared camera,
VOSviewer’s clustering technique. To ensure the optical sensor, or LADAR), a navigation system, and

Figure 14. Keywords analysis.

134 Advances in Structural Engineering 28(1)

GPS. It facilitates in-flight data acquisition and post- The vision-based research in this cluster is not only
flight image processing and can be remotely controlled related to displacement measurement and tracking, but it
by a ground-based navigator. Therefore, the most sig- also opens the door for vibration analysis of structures.
nificant cluster among the research related to drone- Vision-based techniques were used to characterise dy-
based SHM is expected to focus on the development namic properties of structures. The feasibility of using
of the vision-based aspects of the field. The integration drone-captured videos for modal identification of a lab-
of vision-based monitoring with UAVs offers a trans- oratory steel model with correlation functions was ex-
formative approach to SHM. Equipped with high- plored (Yoon et al., 2017) and subsequently utilised the
resolution cameras, UAVs can capture detailed images drone platform for measuring dynamic displacement and
of structures, aiding in the early identification of damage extracting dynamic properties of a footbridge (Vedhus
such as cracks, deformations, or discolouration. The et al., 2019). The modal shapes identified using the
primary advantage of vision-based monitoring is its non- measured displacement from the flying drone platform
contact and remote nature, which allows for the in- correlated well with results obtained from dynamic ac-
spection of hard-to-reach or hazardous areas, such as celerations of contact-type accelerometers demonstrating
high bridges or tall buildings, in a safe and efficient that the error in natural frequencies was less than 1.6%
manner (Lydon et al., 2022). The collected image data compared with the readings from accelerometers. The
can be further analysed using advanced image pro- efficacy of measuring natural frequencies using a UAV
cessing techniques and machine learning algorithms for alongside a stationary camera was evaluated on a steel
automated defect detection and localization. Moreover, frame measuring 9.8 m in length (Chen et al., 2021). The
the combination of UAVs and vision-based monitoring fundamental natural frequency recorded by the UAV,
allows for greater coverage in a shorter time compared to positioned approximately 3 m from the frame, aligned
traditional inspection methods. The methods of vision- with the results from an accelerometer. Further steps have
based SHM fundamentally consist of four stages: cali- been invested in achieving subpixel accuracy through
brating the camera, capturing and correcting the image, interpolation. However, in practical applications of SHM,
measuring the displacement field, and identifying the vibration of structures is typically induced by oper-
damage. The research in this cluster represents a natural ational and ambient loads, such as live loads and tem-
progression of years of camera-based research aimed at perature variations. The vibration amplitude is usually
overcoming challenges related to stability, lighting, and lower compared to that of specimens tested in laborato-
reference images. Imaging-based research was reviewed ries. Additionally, the distance between the cameras and
and summarised in (Sony et al., 2019). A notable ad- the monitored structures is often greater. Consequently,
vantage of UAV sensors is their ability to deliver 3D the vibration amplitude frequently measures less than a
information about structures (Chen et al., 2019), which is pixel in the captured image, significantly compromising
invaluable for large-scale structures and especially ef- measurement accuracy. To take maximum advantage of
fective for monitoring inaccessible areas. They provide vision-based techniques with UAVs in real life, more
superior temporal and spatial resolutions compared to research in this direction should be conducted in the
satellite imagery. Digital Image Correlation (DIC) as a future. The use of thermal images is also a recent ad-
vision-based technique has been widely adopted in UAV- vancement under this cluster. UAVs equipped with in-
related research; however, the minimal movement as- frared thermography capabilities to detect damage on
sociated with high-frequency excitation presents a sig- concrete bridge decks was used for UAV-based SHM
nificant challenge (Lydon et al., 2019a). To enhance the (Omar and Nehdi, 2017). The thermal images captured
precision of deformation measurements, stabilisation were processed using an algorithm that stitched the im-
strategies are essential. Techniques such as 3D point ages together to create a mosaic of the bridge deck, and
cloud and object-based image analysis offer solutions for k-means clustering technique was employed to categorize
measuring displacement at low levels during high- the defects into severity groups. The challenges involved
frequency excitation up to a certain frequency (Chen in this cluster need to be overcome in practical applica-
et al., 2019; Kalaitzakis et al., 2021; Khaloo et al., tions for real advancement of drone-based SHM.
2018b; Kim et al., 2017; Lei et al., 2018, 2020; Pierce
et al., 2018; Rathinam et al., 2008; Sony et al., 2019;
Zhao et al., 2021). However, further investigation is Cluster (2): Enhancing SHM through the integration
required in this domain. Methods for eliminating drone of drones, advanced sensor technologies, and
motion to obtain accurate displacement measurements
require further exploration. Advancing this research
artificial intelligence
direction is essential for effective, applicable drone- The utilisation of drones, combined with advanced
based real SHM. sensor technologies, represents a significant
Fayyad et al. 135

advancement in the field of infrastructure maintenance and facilitated wireless data transfer into cloud-based
and safety. Fibre optic sensors, known for their compact computing systems. This development makes autono-
size and sensitivity, are capable of measuring variations mous monitoring regimes on complex infrastructures
in strain or temperature, thereby potentially delivering feasible if challenges are to be overcome. This fusion of
real-time information about changes in structural be- drones, advanced sensors, and AI heralds a future where
haviour (Lydon et al., 2017; Motwani et al., 2020; Tan structural health monitoring becomes markedly more
et al., 2021, 2024). Consequently, these sensors are efficient, precise, and proactive.
increasingly being used to enhance drones themselves,
with fibres being installed in the wings to develop the
health monitoring of the drones. Moreover, other ad- Cluster (3): Integration of modal analysis, Energy
vanced sensors, such as infrared thermal (IRT) imaging,
harvesting, and deep learning in drone-based SHM
are employed with drones to develop structural health
monitoring. Therefore, the relationships developed in In this cluster, modal analysis serves as the foundation
this cluster primarily stem from advancing the structural for understanding a structure’s dynamic behaviour, as it
health monitoring of the drones themselves, which is identifies the structure’s natural frequencies, damping
indispensable for the development of the field of drone- ratios, and mode shapes. Variations in these parameters
based SHM. In any SHM system, the analysis of mea- can indicate potential structural changes or damage (Li
surement data represents the second crucial step for et al., 2023a), which can be detected by drones equipped
identifying structural characteristics through sensory with appropriate sensors such as accelerometers (Kent
systems (Hassani and Dackermann, 2023). Various data et al., 2023). Measuring the structural dynamic behav-
analysis approaches and algorithms have been developed iour in relation to UAVs is primarily related to vision-
and continue to evolve, influenced by the type of sensors based methods, as discussed in cluster 1. However, in
and the nature of the data collected. The rapidly ex- this cluster, modal analysis pertains to the entire ‘fly-by’
panding field of data science, with its swift advance- process, where ‘fly-by’ refers to a UAV passing over or
ments and innovations in artificial intelligence (AI) and flying around a structure to capture data and images. This
data mining, has brought about significant changes and is used in conjunction with other UAV-based techniques
updates to data analysis methodologies for SHM. While such as photogrammetry, lidar, or infrared imaging (E
conventional data analysis techniques like traditional Polydorou et al., 2021; Poorghasem and Bao, 2022;
signal processing are employed to process datasets and Rathinam et al., 2008; Tse et al., 2023; Wang and Ueda,
evaluate models and hypotheses, AI techniques such as 2023) as well as structure-mounted accelerometers (Na
deep learning are increasingly utilised to reveal under- and Baek, 2017; Sony et al., 2019). Accelerometers can
lying patterns in extensive data sets (Li et al., 2023b). be used in UAV-based SHM in various ways. UAVs can
The integration of AI into these systems significantly wirelessly charge sensors installed on structures and
amplifies the ability to analyse vast volumes of data transmit data from the sensors to the cloud, essentially
collected by sensors via drones. Machine learning al- working as flying sensors. A built-in accelerometer can
gorithms have the potential to detect irregularities and also serve as a sensor for modal analysis if the drone is
anticipate potential structural failures based on identified ‘parked’ on a bridge. The accelerometer will be installed
data patterns. However, their accuracy depends on the inside the drone, which is equipped with a perching
data used to train them, and they can have severe built-in mechanism allowing the drone to adhere to a structure
biases if sufficient data is not available, so more research (Sony et al., 2019). In a study (Na and Baek, 2017),
is needed in this area (Alvarez-Montoya et al., 2020; UAVs were outfitted with a vibration-based non-
Datta et al., 2021; Reddy et al., 2019; Zhou et al., 2022). destructive method to detect damage at an early stage,
AI technologies have increasingly gained prominence, thereby reducing both maintenance expenses and the
attracting considerable interest from the research com- number of sensors affixed to the structure. The proposed
munity in recent years for their potential integration into non-destructive testing (NDT) method employed a
SHM systems, encompassing feature extraction and single piezoelectric material, serving as both exciter and
classification, and applicable in developing intelligent sensor, connected to the UAV through electrical wiring
systems and automation models. This requires the and magnetically attached to ferromagnetic structures or
multiple levels or stages through which information is to a pre-installed magnet on wood or concrete structures.
processed to develop a data-driven model (Cha et al., The method proved effective in identifying various types
2024; Malekloo et al., 2021). Swift progress in cloud- of damage, with one plate exhibiting progressive damage
based computing and wireless technologies, alongside a while the other displayed thickness loss. Additionally, in
trend of decreasing costs for advanced sensors and another investigation (Zhou et al., 2022), the use of an
portable devices, has enabled the deployment of sensors autonomous UAV was proposed for deploying wireless
136 Advances in Structural Engineering 28(1)

sensors in structural monitoring applications. Outdoor which are risky and time-consuming. Advanced robotic
experiments validated the vision-aided control of the technologies offer significant potential for automating
UAV for precise sensor placement, demonstrating the defect inspections and vibration measurements of
UAV’s ability to land within a 10 cm radius of a pre- structures i.e. bridges (Smarsly et al., 2023). The inte-
defined point. This approach underscores the potential gration of automation and UAV-based SHM is a an
for enhancing the efficiency and accuracy of structural emerging area of research. Here, the convergence of
monitoring through advanced UAV technology coupled UAVs and robotics gives rise to automated and intelli-
with modal analysis. Meanwhile, energy harvesting gent SHM systems, capable of performing comprehen-
plays a crucial role in sustaining the operation of the sive assessments of structures with minimal human
SHM system, particularly the drone and its embedded intervention. These automated systems, powered by
sensors. By converting ambient energy, often sourced advanced robotic technologies, can navigate through
from the structural system’s vibrations or solar energy, intricate environments, collect vital data, and identify
into electrical energy, the SHM system can operate in- potential structural concerns (Sony et al., 2019). Yet,
dependently of external power sources. This enhances most research on the application of robotic machines in
the system’s sustainability and longevity, enabling health monitoring and inspection fields is related to
continuous, long-term monitoring. drones, with less focus on the application of mobile
One of the foremost data processing challenges associated robots. This area requires further investigation to achieve
with UAV-based SHM, as discussed in cluster 2, involves effective SHM. It could be enhanced by the capabilities
managing the substantial volumes of data that can be pro- of Building Information Modelling (BIM), which pro-
duced. UAVs are capable of capturing high-resolution images, vides a detailed, three-dimensional, information-rich
videos, dynamic behaviour parameters, and other data swiftly, model of a structure. These technologies allow the
often resulting in large datasets that necessitate processing and data collected by UAVs and robots to be contextualised,
analysis. This demands considerable storage and computing visualised, and analysed more efficiently. However, a
resources, as well as specialised software tools for data challenge faced is that BIM models do not currently
management and analysis. Another challenge concerns data achieve the accuracy needed for SHM, such as sub-
accuracy and consistency. UAVs can capture data from millimetre for deflections and micro-strain for strain
various perspectives and at different times, potentially leading measurements. Thus, the fusion of automation, robotics,
to inconsistencies in data quality and accuracy. Furthermore, UAVs, and BIM can potentially revolutionise the field of
data may require correction for factors such as camera dis- SHM, leading to more accurate, efficient, and safer
tortion or sensor errors, which can additionally affect data structure inspections (Perry et al., 2020; Poorghasem and
accuracy (Martinez et al., 2020). Deep learning techniques Bao, 2022; Wang and Ueda, 2023), but this will not
add another layer of sophistication to the UAV-based SHM happen until the above mentioned challenges are solved.
system under this cluster. These algorithms can process and The main benefits of robots are to overcome the
analyse the large amounts of data collected by drones in a limitations of UAVs, such as restricted flight times and
highly efficient manner, identifying complex patterns and payload capacities. Unmanned ground vehicles (UGVs)
detecting anomalies that may indicate structural damage. By and hybrid robots show promise in enhancing the per-
training these algorithms with labelled datasets, they can formance of vibration measurements. Collaborating
accurately classify different types of structural damage and UGVs with unmanned surface vehicles (USVs) and
potentially predict future structural failures. This fusion of UAVs can enhance the accuracy of vibration measure-
modal analysis, energy harvesting, and deep learning creates ments through the improved precision of robot navi-
an advanced, self-sustaining, and highly accurate drone-based gation. The navigation systems of UGVs can assist in
SHM system that can significantly enhance the maintenance precisely locating UAVs. Additionally, using a tri-rotor
and safety of various infrastructures (Agyemang et al., 2022; hybrid UAV is promising for vibration measurement due
Chen et al., 2017; Dong et al., 2020; Javadinasab to its lower power consumption (Tian et al., 2022). It is
Hormozabad et al., 2021; Kang and Cha, 2018; Kao et al., noted from the Overlay visualisation Figure 15 (to be
2023; Kulkarni et al., 2023; Lydon et al., 2019b; Perez et al., explained in the next section) that this cluster involves
2019; Ye et al., 2022). papers presenting the most recent research. This high-
lights the significant role of automation and robotics in
advancing the capabilities of drone-Based SHM.
Cluster (4): automation and robotics in The four distinct clusters coalesce under the umbrella
of UAV-based SHM where different techniques can be
drone-based SHM
integrated and various solutions can be harmonised. One
Traditional methods of defect detection rely heavily on of the primary challenges in the field of SHM is that no
specialised inspection vehicles and visual inspections, single technology suffices for comprehensive
Fayyad et al. 137

Figure 15. Keywords timeline in UAV-based SHM.

monitoring, necessitating the integration of various algorithms and software that can seamlessly integrate
technologies. The complexity and variability of struc- and analyse data from multiple SHM technologies.
tures mean that different damage mechanisms may not be
effectively captured by a singular method. For instance,
while vibration-based techniques are adept at identifying
changes in structural dynamics, they may not pinpoint
Research timeline
the exact location of minor damages as effectively as Figure 15 shows that the progress of vision UAV-based
vision-based techniques. Consequently, merging SHM over the past decade. Notably, a concentration of
vibration-based methods with vision-based technologies research during 2016-2018 focused on “Photogrammetry”
can offer a more holistic view of a structure’s health. and “digital image correlation”, centring around well-
Moreover, the transition of theoretical models into real- established techniques. The span of 2018-2020 shows
world applications presents another significant barrier. the emergence of keywords such as “computer vision”,
These models often require adjustments to accurately “3D scanning”, “3D reconstruction”, “remote sensing”,
reflect the diverse behaviours of structures under dif- and “machine learning”. More recent research topics, in-
ferent stages of damage, which can vary greatly de- cluding “robotics”, “artificial intelligence”, “automation”,
pending on materials, design, and environmental and “digital twin”, suggest a shift in the research focus in
conditions. This underscores the need for adaptive and this area. While early contributions prioritised data col-
multidisciplinary approaches in SHM, capable of in- lection for inspection purposes, later publications aimed to
terpreting complex data from varied sources to accu- leverage “artificial intelligence” and “robotics” to automate
rately assess structural integrity. Additionally, SHM. This shift might be attributed to the novel per-
challenges such as sensor placement optimization, data spectives and methods brought about by advancements in
management, and the interpretation of vast amounts of automation and robotics. “Digital twin” is a relatively new
data from different technologies into actionable insights research topic with limited coverage so far (Yoon et al.,
further complicate the SHM landscape. Addressing these 2022). “Deep learning” and “machine learning” have been
challenges requires ongoing research, interdisciplinary integrated with other keywords in recent years, signifying
collaboration, and the development of advanced their growing relevance in the field.
138 Advances in Structural Engineering 28(1)

Reviews in the field techniques is crucial. These should aim to mitigate

the effects of drone vibration and movement, en-
This paper presents a quantitative analysis and review of abling more precise measurements from the
the field of SHM using drones to provide a far more in collected data.
depth understanding of the topic relating drone use for - Continued advancements in image processing, in-
SHM using scientometric analysis. The scientometric cluding the use of machine learning and deep
research results for SHM and UAV show a nine publica- learning algorithms, are required to handle the vast
tions named as review papers; four of these nine papers amounts of data generated by drones more effec-
were excluded for not being related specifically to UAV- tively. These algorithms can enhance the ability to
SHM, however they mention SHM and damage detection detect subtle changes and anomalies in structures,
as a possible applications with drones. Further to the ad- thereby predicting potential failures more accurately.
ditional manual check using more traditional techniques - Combining data from various sensors, including
(such as Google scholar), two more papers were added thermal, optical, and structural sensors, can provide a
which are very recent papers. The content analysis of the more comprehensive understanding of structural
review papers is presented in Table 4. The UAV-based health. Research into effective methods for inte-
SHM is a very specific field and also is a very recent one but grating and analysing data from these diverse sources
it intersects with many research topics and these review could lead to more robust SHM systems.
papers covers/overlaps in different ways with the fly-by - Leveraging artificial intelligence further to develop
technique. predictive analytics for SHM could revolutionise the
As summarised in Table 4, the existing review papers field. By employing AI algorithms that analyse
cover various aspects, fields, and applications that could historical data and real-time inputs, it’s possible to
intersect, to varying extents, with drone-based SHM forecast potential structural issues before they be-
“fly-by”. To the author’s knowledge, there is no review come critical.
paper that is meant for drone-based SHM field either - The development of fully autonomous drones that
traditional qualitative review nor scientometric review. can perform regular monitoring without human in-
This paper, positioned at the core of UAV-based SHM, in tervention represents a significant step forward. This
conjunction with other reviews that overlap to some would include the capability for drones to navigate
extent with field, offers a new quantitative insights into and operate in complex environments autonomously,
the advancements in this field. Furthermore, while the ensuring continuous monitoring and immediate
review publications primarily adopt a qualitative ap- response.
proach, reflecting the authors’ comprehension and per- - As the precision of measurements is critical for the
spectives on the field, this paper presents a quantitative early detection of structural issues, future research
analysis of its current state. should also focus on achieving subpixel accuracy in
By combining both quantitative and qualitative ana- displacement measurements. This involves refining
lyses, a more comprehensive overview of the field is ob- the resolution of image data beyond the current ca-
tained with new insights into logical clusters of research pabilities, allowing for the detection of minute
activities. This allows for a clearer picture of potential changes.
directions for future progression, mitigating any inherent - While some of the novel techniques have been ap-
subjectivity. plied to UAV-based SHM, many have predominantly
been examined under research conditions. Therefore,
they require further exploration in real-life envi-
Future directions ronments to fully assess and understand their benefits
Building upon the foundations established in drone-based and challenges.
SHM, several directions for future research have been - Digital twin technology represents a cutting-edge
identified to advance this field significantly. technology that bridges the physical and virtual
worlds. Further research is required in drone-based
- The significant power consumption of different ex- SHM to fully exploit the capabilities of digital twin
isting technologies pose a challenge. There is a need technology.
for further advancements in self-powered sensors.
Novel vision-based sensors must be insensitive to These directions not only highlight the potential for
light conditions, necessitating the development of technological advancements but also underscore the need
enhanced algorithms for image processing. Also, to for a multidisciplinary approach involving engineering,
improve the accuracy of vision-based monitoring data science, robotics, and regulatory studies to fully realise
systems, further development of stabilisation the benefits of drone-based SHM. Further research and
Fayyad et al. 139

Table 4. Review Papers in Fields That Intersect With Fly-By SHM

Ref Year Scope Focus/Contribution Limits/Coverage

(Sony et al., 2019 A literature review of next- - First review paper that presents a review - The paper covers all emergent
2019) generation smart sensing of UAVs for SHM applications. smart sensing technologies
technology in SHM - It presents a review on UAVs in the - Only covers up to 2019
context of other emergent smart
sensing technologies such as cameras,
drones, robotic sensors.
- It covers the development of the first
decade in the field with a round
30 publications.
(Tian et al., 2022 Applications and future trends - It provides a review on existing robotic - Covers different robotic
2022) of intelligent robotic systems systems in three areas; inspections of systems such as mobile
for SHM structural defects, dynamic response robots, wall-climbing robots,
and SHM potential. cable-climbing robots, and
flying UAVs.
- Many multimodal robots were reviewed - Presents a review of the
for their potential application in drones with the focus on
automating SHM tasks of bridges. integrating the technology
with various NDT devices
and advanced sensors such as
synthetic aperture radar
(SAR), active microwave
imaging system and robotic
arms etc. for a versatile flying
drone.
- The review was focused on
bridge SHM.
(Zinno et al., 2023 AI approaches and new - Highlights how AI, especially deep - Reviews the application of
2023) technologies in bridges SHM learning, and some technologies such as different new technologies
drone technology and 3D printers could for SHM such as AI, UAVs
be used to improve SHM of bridges. and 3D printing.
- The paper presents the applications of - Review is focused on bridge
optimisation algorithms for optimising SHM.
sensors for SHM.
(Hassani and 2023 Review of advanced sensor - Provides a review on sensing techniques - Techniques are reviewed in
Dackermann, technologies for NDT and such as fibre optics, laser vibrometry, the context of providing
2023) SHM acoustic emission, ultrasonics, input parameters for NDT.
thermography, drones, - The UAVs are presented as a
microelectromechanical systems part of the techniques, so the
(MEMS), magnetostrictive sensors, and focus is not only the UAV-
next-generation technologies. based SHM.
(Poorghasem 2023 Review of robot-based - Discusses robot-based hardware, the - Covers different robotic
and Bao, automated measurement of machine vision approaches, and the platforms, so the focus is not
2022) vibration for civil engineering capabilities of automated measurement only the UAV
structures of structural vibration. - Different applications are
covered for UAVs (bridges,
towers, buildings, wind
turbines)

(continued)
140 Advances in Structural Engineering 28(1)

Table 4. (continued)

Ref Year Scope Focus/Contribution Limits/Coverage

(Liang et al., 2023 Advancement, challenges, and -Outlines types of UAVs - Main focus of the review is the
2023) future directions of UAVs for construction industry either
monitoring and inspection in the types or the related
technologies.
- Technologies related to UAVs
- Limitations & challenges - Highlights some limitations
and challenges that are
common in UAV based fields.
Future directional insights
(Wang and 2023 A review study on unmanned - Presents UAVs in different field - The focus of the paper is
Ueda, 2023) aerial vehicle and mobile applications for inspection and damage mainly damage inspection in
robot technologies on detection reinforced concrete
damage inspection of structures.
reinforced concrete - Discusses building damage inspection, - Covers mobile robot
structures challenges & future directions. application in damage
inspection.

collaboration across these domains will be crucial in relatively new research area with limited coverage,
overcoming the existing challenges and unlocking the full suggesting a potential direction for future studies. One of
potential of drones in the field. the primary challenges in the field of SHM is that no
single technology suffices for comprehensive monitor-
ing, necessitating the integration of various technolo-
Conclusion gies. The complexity and variability of structures mean
that different damage mechanisms may not be effectively
Structural Health Monitoring using Drones has the po- captured by a singular method. To address these chal-
tential to be a highly effective approach (fly-by technique). lenges, ongoing research, interdisciplinary collabora-
A mixed review approach combining scientometric and tion, and the development of advanced algorithms and
qualitative methodologies has revealed a general rise in software that can seamlessly integrate and analyse data
publications covering research in SHM using UAVs since from multiple SHM technologies are needed.
2010, with noticeable surges in 2017-2018 and 2021-2022.
This study predicts continued growth in this research field.
Keyword analysis has revealed four main clusters of Acknowledgments
research within the field of fly-by SHM. These are (1) the This research acknowledges the University of Melbourne’s “FEIT
application of UAV-enabled vision-based monitoring, (2) Visiting Fellow Scheme” for bringing together the two institutions
the integration of drones, advanced sensor technologies, “Queen’s University of Belfast” and the “University of Mel-
and artificial intelligence, (3) drone-based SHM integrating bourne” to develop the research. The authors acknowledge
modal analysis, energy harvesting, and deep learning and Professors Priyan Mendis and Lu Aye of Dept of Infrastructure
lastly, and (4) automation and robotics in drone- Engineering, University of Melbourne for their valuable advice on
based SHM. this paper.
A shift in research focus was identified from data
collection towards leveraging artificial intelligence and
robotics to automate SHM over the last years. The in- Declaration of conflicting interests
tegration of “deep learning” and “machine learning” The author(s) declared no potential conflicts of interest with re-
with other keywords signifies their growing relevance in spect to the research, authorship, and/or publication of this article.
the field. Some gaps and future research suggestions are:
(i) further research and development to enhance the
capabilities of drones for SHM, given the potential Funding
growth of the UAV market in this field, (ii) shift towards The author(s) disclosed receipt of the following financial support
using AI and robotics to automate SHM The creation of for the research, authorship, and/or publication of this article: This
smart, autonomous drones, that can conduct inspections research was funded by Daphne Jackson Trust, the Royal Society
alongside processing SHM sensor data is a key area for and the Engineering and Physical Sciences Research Council
development, (iii) the use of “digital twins” in SHM is a (EPSRC) https://daphnejackson.org/fellow/dr-tahreer-fayyad/.
Fayyad et al. 141

ORCID iDs Dong C-Z, Celik O, Catbas FN, et al. (2020) Structural dis-
placement monitoring using deep learning-based full field
Tahreer M. Fayyad  https://orcid.org/0000-0001-6978-6134
optical flow methods. Structure and Infrastructure Engi-
Kun Feng  https://orcid.org/0000-0003-1110-3681
neering 16(1): 51–71.
Efstathios P, Robinson D, Taylor S, et al. (2021) Health moni-
References
toring of structures using integrated unmanned aerial vehi-
Adams DE (2007) Health Monitoring of Structural Materials and cles (UAVs). International Workshop on Civil Structural
Components: Methods with Applications. Chichester, En- Health Monitoring 2021, 243–255.
gland: John Wiley & Sons, Ltd. Ellenberg A, Branco L, Krick A, et al. (2015) Use of unmanned
Agyemang IO, Zhang X, Acheampong D, et al. (2022) Auton- aerial vehicle for quantitative infrastructure evaluation.
omous health assessment of civil infrastructure using deep Journal of Infrastructure Systems 21(3): 4014054.
learning and smart devices. Automation in Construction 141: Feng K, Casero M and González A (2019) The use of acceler-
104396. ometers in UAVs for bridge health monitoring. In: Pro-
Alvarez-Montoya J, Carvajal-Castrillón A and Sierra-Pérez J ceedings of the 13th International Conference on
(2020) In-flight and wireless damage detection in a UAV Applications of Statistics and Probability in Civil Engi-
composite wing using fiber optic sensors and strain field neering (ICASP 2019). Seoul National University.
pattern recognition. Mechanical Systems and Signal Pro- Hassani S and Dackermann U (2023) A systematic review of
cessing 136: 106526. advanced sensor technologies for non-destructive testing and
Bao Y, Chen Z, Wei S, et al. (2019) The state of the art of data structural health monitoring. Sensors 23.
science and engineering in structural health monitoring. He Q, Wang G, Luo L, et al. (2017) Mapping the managerial areas
Engineering 5(2): 234–242. of Building Information Modeling (BIM) using sciento-
Calvi GM, Moratti M, O’Reilly GJ, et al. (2019) Once upon a time metric analysis. International Journal of Project Manage-
in Italy: the tale of the Morandi bridge. Structural Engi- ment 35(4): 670–685.
neering International 29(2): 198–217. Hossain T, Segura S and Okeil AM (2020) Structural effects of
Cha Y-J, Ali R, Lewis J, et al. (2024) Deep learning-based temperature gradient on a continuous prestressed concrete
structural health monitoring. Automation in Construction girder bridge: analysis and field measurements. Structure
161: 105328. and Infrastructure Engineering 16(11): 1539–1550.
Chen S, Laefer DF and Mangina E (2016) State of technology Li H, Ou J, Zhao X, et al. (2006) Structural health monitoring
review of civilian UAVs. Recent Patents on Engineering system for the shandong binzhou yellow river highway
10(3): 160–174. bridge. Computer-Aided Civil and Infrastructure Engi-
Chen X, Kopsaftopoulos F, Cao H, et al. (2017) Intelligent flight neering 21(4): 306–317.
state identification of a self-sensing wing through neural Ivancheva L (2008) Scientometrics today: a methodological
network modelling. In Proceedings of the 11th International overview. COLLNET Journal of Scientometrics and Infor-
Workshop on Structural Health Monitoring. IWSHM. mation Management 2(2): 47–56.
Chen S, Laefer DF, Mangina E, et al. (2019) UAV bridge in- Javadinasab Hormozabad S, Gutierrez Soto M and Adeli H (2021)
spection through evaluated 3D reconstructions. Journal of Integrating structural control, health monitoring, and energy
Bridge Engineering 24(4). harvesting for smart cities. Expert Systems 38(8): e12845.
Chen G, Liang Q, Zhong W, et al. (2021) Homography-based Kalaitzakis M, Vitzilaios N, Rizos DC, et al. (2021) Drone-based
measurement of bridge vibration using UAV and DIC StereoDIC: system development, experimental validation
method. Measurement 170: 108683. and infrastructure application. Experimental Mechanics 61:
Darko A, Chan APC, Adabre MA, et al. (2020) Artificial intel- 981–996.
ligence in the AEC industry: scientometric analysis and Kang D and Cha Y (2018) Autonomous UAVs for structural
visualization of research activities. Automation in Con- health monitoring using deep learning and an ultrasonic
struction 112: 103081. beacon system with geo-tagging. Computer-Aided Civil and
Datta A, Augustin MJ, Gaddikeri KM, et al. (2021) Damage Infrastructure Engineering 33(10): 885–902.
detection in composite aircraft wing-like test-box using Kao S-P, Chang Y-C and Wang F-L (2023) Combining the
distributed fiber optic sensors. Optical Fiber Technology 66: YOLOv4 deep learning model with UAV imagery pro-
102651. cessing technology in the extraction and quantization of
Dennison PE, Brewer SC, Arnold JD, et al. (2014) Large wildfire cracks in bridges. Sensors 23(5): 2572.
trends in the western United States, 1984–2011. Geophysical Kent C, Bunce A, O’Higgins C, et al. (2023) Uncertainty in bridge
Research Letters 41(8): 2928–2933. mode shapes based on accelerometer location. In: 14th In-
Ding X and Yang Z (2022) Knowledge mapping of platform ternational Conference on Applications of Statistics and
research: a visual analysis using VOSviewer and CiteSpace. Probability in Civil Engineering (ICASP 2023): Proceed-
Electronic Commerce Research 22(3): 787–809. ings, Dublin, Ireland.
142 Advances in Structural Engineering 28(1)

Khaloo A, Lattanzi D, Cunningham K, et al. (2018a) Unmanned low cost wireless action camera. Mechanical Systems and
aerial vehicle inspection of the Placer River Trail Bridge Signal Processing 121: 343–358.
through image-based 3D modelling. Structure and Infra- Lydon D, Taylor SE, Lydon M, et al. (2019b) Development and
structure Engineering 14(1): 124–136. testing of a composite system for bridge health monitoring
Khaloo A, Lattanzi D, Jachimowicz A, et al. (2018b) Utilizing utilising computer vision and deep learning. Smart Struc-
UAV and 3D computer vision for visual inspection of a large tures and Systems 24(6): 723–732.
gravity dam. Frontiers in Built Environment 4: 31. Lydon D, Kromanis R, Lydon M, et al. (2022) Use of a roving
Khan SM, Atamturktur S, Chowdhury M, et al. (2016) Integration computer vision system to compare anomaly detection
of structural health monitoring and intelligent transportation techniques for health monitoring of bridges. Journal of Civil
systems for bridge condition assessment: current status and Structural Health Monitoring 12(6): 1299–1316.
future direction. IEEE Transactions on Intelligent Trans- Malekloo A, Ozer E, AlHamaydeh M, et al. (2021) Machine
portation Systems 17(8): 2107–2122. learning and structural health monitoring overview with
Kim H, Lee J, Ahn E, et al. (2017) Concrete crack identification emerging technology and high-dimensional data source
using a UAV incorporating hybrid image processing. Sensors highlights. Structural Health Monitoring 21(4): 1906–1955.
17(9), 2052. Martinez P, Al-Hussein M and Ahmad R (2019) A scientometric
Ko JM and Ni YQ (2005) Technology developments in structural analysis and critical review of computer vision applications
health monitoring of large-scale bridges. Engineering for construction. Automation in Construction 107: 102947.
Structures 27(12): 1715–1725. Martinez JG, Gheisari M and Alarcón LF (2020) UAV integration in
Kressel I, Dorfman B, Botsev Y, et al. (2015) Flight validation of current construction safety planning and monitoring processes:
an embedded structural health monitoring system for an case study of a high-rise building construction project in Chile.
unmanned aerial vehicle. Smart Materials and Structures Journal of Management in Engineering 36(3).
24(7): 75022. Mingers J and Leydesdorff L (2015) A review of theory and
Kulkarni NN, Raisi K, Valente NA, et al. (2023) Deep learning practice in scientometrics. European Journal of Operational
augmented infrared thermography for unmanned aerial ve- Research 246(1): 1–19.
hicles structural health monitoring of roadways. Automation Mohsan SA, Khan MA, Noor F, et al. (2022) Towards the un-
in Construction 148: 104784. manned aerial vehicles (UAVs): a comprehensive review.
Lei B, Wang N, Xu P, et al. (2018) New crack detection method Drones 6(147).
for bridge inspection using UAV incorporating image pro- Mongeon P and Paul-Hus A (2016) The journal coverage of Web
cessing. Journal of Aerospace Engineering 31(5): 4018058. of Science and Scopus: a comparative analysis. Sciento-
Lei B, Ren Y, Wang N, et al. (2020) Design of a new low-cost metrics 106: 213–228.
unmanned aerial vehicle and vision-based concrete crack Motwani P, Perogamvros N, Taylor S, et al. (2020) Experimental
inspection method. Structural Health Monitoring 19(6): investigation of strain sensitivity for surface bonded fibre
1871–1883. optic sensors. Sensors and Actuators A: Physical 303:
Li Z, Lin W and Zhang Y (2023a) Drive-by bridge damage 111833.
detection using Mel-frequency cepstral coefficients and Na WS and Baek J (2017) Impedance-based non-destructive
support vector machine. Structural Health Monitoring 22(5): testing method combined with unmanned aerial vehicle
3302–3319. for structural health monitoring of civil infrastructures.
Li Z, Lin W and Zhang Y (2023b) Real-time drive-by bridge Applied Sciences 7(1): 15.
damage detection using deep auto-encoder. Structures 47: Noel AB, Abdaoui A, Elfouly T, et al. (2017) Structural health
1167–1181. monitoring using wireless sensor networks: a comprehensive
Liang H, Lee S-C, Bae W, et al. (2023) Towards UAVs in survey. IEEE Communications Surveys & Tutorials 19(3):
construction: advancements, challenges, and future direc- 1403–1423.
tions for monitoring and inspection. Drones 7(3): 202. Nooralishahi P, Ibarra-Castanedo C, Deane S, et al. (2021) Drone-
Lydon M, Taylor S, Sivakumar V, et al. (2012) Structural health based non-destructive inspection of industrial sites: a review
monitoring of novel concrete arch system. Bridge and and case studies. Drones 5(4): 106.
Concrete Research in Ireland conference proceedings. Ire- Olawumi TO and Chan DWM (2018) A scientometric review of
land: Dublin Institute of Technology and Trinity College global research on sustainability and sustainable develop-
Dublin. ment. Journal of Cleaner Production 183: 231–250.
Lydon M, Taylor SE, Doherty C, et al. (2017) Bridge weigh-in- Omar T and Nehdi ML (2017) Remote sensing of concrete bridge
motion using fibre optic sensors. In: Proceedings of the decks using unmanned aerial vehicle infrared thermography.
Institution of Civil Engineers - Bridge Engineering. ICE Automation in Construction 83: 360–371.
Publishing, Vol. 170. 219–231. Ozer E and Feng MQ (2019) Structural reliability estimation with
Lydon D, Lydon M, Taylor S, et al. (2019a) Development and participatory sensing and mobile cyber-physical structural
field testing of a vision-based displacement system using a health monitoring systems. Applied Sciences 9(14): 2840.
Fayyad et al. 143

Pedram M, Taylor SE, Robinson D, et al. (2021) Subsurface monitoring. Structural Control and Health Monitoring
defect detection in concrete using infrared thermography. In: 26(3): e2321.
10th International Conference on Structural Health Moni- Tan X, Fan L, Huang Y, et al. (2021) Detection, visualization,
toring of Intelligent Infrastructure: Advanced Research and quantification, and warning of pipe corrosion using dis-
Real World Applications (SHMII-10), International Society tributed fiber optic sensors. Automation in Construction 132:
for Structural Health Monitoring of Intelligent …, 103953.
pp. 1155–1160. Tan X, Poorghasem S, Huang Y, et al. (2024) Monitoring of
Pedram M, Taylor S, Hamill G, et al. (2024) Objective charac- pipelines subjected to interactive bending and dent using
terisation of reinforced concrete with progressive corrosion distributed fiber optic sensors. Automation in Construction
defects through clustering and thresholding of infrared im- 160: 105306.
ages. Measurement 225: 114017. Tian Y, Chen C, Sagoe-Crentsil K, et al. (2022) Intelligent robotic
Perez H, Tah JHM and Mosavi A (2019) Deep learning for de- systems for structural health monitoring: applications and
tecting building defects using convolutional neural net- future trends. Automation in Construction 139: 104273.
works. Sensors 19(16): 3556. Tokognon AC, Gao B, Tian GY, et al. (2017) Structural health
Perry BJ, Guo Y, Atadero R, et al. (2020) Streamlined bridge monitoring framework based on Internet of Things: a survey.
inspection system utilizing unmanned aerial vehicles IEEE Internet of Things Journal 4(3): 619–635.
(UAVs) and machine learning. Measurement 164: Tse K-W, Pi R, Sun Y, et al. (2023) A novel real-time autonomous
108048. crack inspection system based on unmanned aerial vehicles.
Pierce SG, Burnham K, McDonald L, et al. (2018) Quantitative Sensors 23(7): 3418.
inspection of wind turbine blades using UAV deployed Van Eck NJ and Waltman L (2014) VOSviewer Manual. Neth-
photogrammetry. In: 9th European Workshop on Structural erlands: Univeristeit Leiden.
Health Monitoring (EWHM 2018). Manchester, UK, 10–13, van Eck NJ and Waltman L (2017) Citation-based clustering of
July, 2018. publications using CitNetExplorer and VOSviewer. Scien-
Polydorou E, Taylor S, Robinson D, et al. (2021) Structural health tometrics 111(2): 1053–1070.
monitoring of structures using UAVs. In: 10th International Vedhus H, Jong-Woong P, Hyungchul Y, et al. (2019) Vision-
Conference on Structural Health Monitoring of Intelligent based modal survey of civil infrastructure using unmanned
Infrastructure: Advanced Research and Real World Appli- aerial vehicles. Journal of Structural Engineering 145(7):
cations (SHMII-10). International Society for Structural 04019062.
Health Monitoring of Intelligent …, pp. 1161–1165. Wang J and Ueda T (2023) A review study on unmanned aerial
Poorghasem S and Bao Y (2022) Review of robot-based auto- vehicle and mobile robot technologies on damage inspection
mated measurement of vibration for civil engineering of reinforced concrete structures. Structural Concrete 24(1):
structures. Measurement 207: 112382. 536–562.
Rathinam S, Kim ZW and Sengupta R (2008) Vision-based Xu J, Dong Y, Zhang Z, et al. (2016) Full scale strain monitoring
monitoring of locally linear structures using an unmanned of a suspension bridge using high performance distributed
aerial vehicle. Journal of Infrastructure Systems 14(1). fiber optic sensors. Measurement Science and Technology
52–63. 27(12): 124017.
Reagan D, Sabato A and Niezrecki C (2018) Feasibility of using Ye X, Ma S, Liu Z, et al. (2022) Post-earthquake damage rec-
digital image correlation for unmanned aerial vehicle ognition and condition assessment of bridges using UAV
structural health monitoring of bridges. Structural Health integrated with deep learning approach. Structural Control
Monitoring 17(5): 1056–1072. and Health Monitoring 29(12): e3128.
Reddy A, Indragandhi V, Ravi L, et al. (2019) Detection of Cracks Yoon H, Hoskere V, Park J-W, et al. (2017) Cross-correlation-
and damage in wind turbine blades using artificial based structural system identification using unmanned aerial
intelligence-based image analytics. Measurement 147: vehicles. Sensors 17(9).
106823. Yoon S, Lee S, Kye S, et al. (2022) Seismic fragility analysis of
Sarmadi H, Entezami A, Yuen K-V, et al. (2023) Review on deteriorated bridge structures employing a UAV inspection-
smartphone sensing technology for structural health moni- based updated digital twin. Structural and Multidisciplinary
toring. Measurement 223: 113716. Optimization 65(12): 346.
Smarsly K, Dragos K, Stührenberg J, et al. (2023) Structural Zhang G, Liu Y, Liu J, et al. (2022) Causes and statistical
health monitoring of civil infrastructure using mobile robots. characteristics of bridge failures: a review. Journal of Traffic
In: Skatulla S and Beushausen H (ed) Advances in Infor- and Transportation Engineering 9(3): 388–406.
mation Technology in Civil and Building Engineering: Zhao S, Kang F, Li J, et al. (2021) Structural health monitoring
Cham: Springer Nature Switzerland, pp. 127–138. and inspection of dams based on UAV photogrammetry with
Sony S, Laventure S and Sadhu A (2019) A literature review of image 3D reconstruction. Automation in Construction 130:
next-generation smart sensing technology in structural health 103832.
144 Advances in Structural Engineering 28(1)

Zhou G-D and Yi T-H (2013) Thermal load in large-scale bridges: Zink J and Lovelace B (2015) Unmanned Aerial Vehicle Bridge
a state-of-the-art review. International Journal of Distrib- Inspection Demonstration Project. St. Paul, MN: Minnesota
uted Sensor Networks 9: 217983. Department of Transportation, Research Services & Library.
Zhou H, Lynch J and Zekkos D (2022) Autonomous wireless Zinno R, Haghshenas SS, Guido G, et al. (2023) The state of the
sensor deployment with unmanned aerial vehicles for art of artificial intelligence approaches and new technologies
structural health monitoring applications. Structural Control in structural health monitoring of bridges. Applied Sciences
and Health Monitoring 29(6): e2942. 13(1): 97.
Zhu XQ and Law SS (2015) Structural health monitoring based on Zonta D, Glisic B and Adriaenssens S (2014) Value of infor-
vehicle-bridge interaction: accomplishments and challenges. mation: impact of monitoring on decision-making. Struc-
Advances in Structural Engineering 18(12): 1999–2015. tural Control and Health Monitoring 21(7): 1043–1056.