Accelerating Power Grid Monitoring with

Flying Robots and Artificial Intelligence

A. Chehri, Senior Member, IEEE, G. Jeon, Senior Member, IEEE, I. Fofana, Senior Member,
IEEE, A. Imran, Senior Member, IEEE, and R. Saadane, Member, IEEE

The miniaturization of optical sensors, the era of digital

Abstract—The digital revolution is expected to surpass all imaging, and advances in image processing, driven by the
previous economic transformations in scale, scope, and excellent computing power available today, promise the drone
complexity. Digital technologies are making electrical grids more mapping sector many perspectives. After being born as flying
connected, more reliable, and sustainable. Several efforts have
been made to revamp the electric grid and modernize century-old toys, and after becoming robotic cameras, drones today carry
systems. In the near future, we expect to see a revolution in power enough intelligence to assist humans in complete autonomy.
grind monitoring with incredible results through artificial Thanks to the versatility and high mobility of UAVs, low-
intelligence, big data, drones, among others. The business altitude UAVs are extensively used in diverse fields for
opportunity for using drones in the energy sector is impressive, different applications and purposes. Drones can be used to
although very few companies have joined in its implementation. monitor electrical networks in real-time and broadcast live
The drone allows the safe remote overflight of high voltage power
lines. They can be deployed to detect, inspect, and diagnose the detailed high-resolution shots of the various production entities,
defects of the power line infrastructure. In this paper, we provide the transmission, transport, and distribution networks. More
a state-of-art of using drones in smart grid monitoring. We secure and less expensive than human intervention, drones can
demonstrate and propose an architecture based on Faster R-CNN replace the inspection procedures traditionally performed by
for detecting ice accretion on power lines. Finally, we shed light on humans [3]. Their use increases mission safety, as inspections
opportunities and future trends of these emerging technologies of energy assets often require employees and contractors to
that can guide future research directions.
Index Terms—Smart Grid, Drone, Unmanned Aerial Vehicles work at heights. A drone with a thermal image camera can
Artificial Intelligence, Power Grid Monitoring, Solar Plant. detect overheating parts or places requiring maintenance. This
analysis makes it possible to identify, prepare, and secure an
I. INTRODUCTION intervention on many infrastructure types, such as electricity
Drones, "the new flying IoT," or "flying robots," have pylons, wind farms, or photovoltaic panels [4]-[5].
become essential tools in many areas. Also known by their Artificial intelligence (AI) has recently flourished in
acronym, UAV (Autonomous Unmanned Vehicles), they can different applications through advanced machine learning, deep
perform an unlimited number of tasks. learning, and reinforcement learning. Drones, combined with
Technically, the drone consists of three segments (1) The air artificial intelligence, will play a crucial role in various real-
segment includes one or more air vectors carrying the payload, time monitoring applications.
such as a camera, (2) The ground segment comprises one or All industries can benefit from the technological
more control and information collection stations, the support developments that result, including drones. We show that the
team, and (3) the wireless component is a set of data links classifier based on Faster R-Convolutional neural networks
connecting the ground and air segments [1]. (CNN) gives better results than the more classical methods and
The UAV is an aircraft without a human pilot on board. This can be used for Ice accretion detection on power cables.
autonomy is conferred on these flying robots by the navigation This paper defines drones and their addressable marketplace
instruments onboard the drone: GPS, inertial station, compass, for smart grid monitoring. Two example case studies are
barometer, etc. Drones can play an active role when they are provided, UAV-based solar inspections and ice detection on
equipped with a payload to interact with their environment. electrical power cables. For the last application, we applied a
However, the vast majority of drones play an observer role, and deep learning architecture, i.e., Faster-RCNN, with ResNet-101
their payload is a sensor such as a simple camera. as a backbone network. To the best of our knowledge, Faster-
Beyond applications leisure and small consumer toys, the RCNN with ResNet-101 architecture is used for the first time
primary uses of the unmanned platforms onboard are for the ice accretion on electric power lines detection. The
designated by the acronym “5D” for Dull, Dirty, Dangerous, images are collected from different online available data
Dear, Difficult; either repetitive, compromising, dangerous, at sources for testing and training of the model. The real-time
a high cost and difficult. Although using a drone allows humans image collected from the drone can directly be fed deep learning
to stay away from danger and take turns throughout the mission, architecture for ice detection. The framework is capable of
the management of such a platform is a complex task that assisting Utility companies to remote monitor their
requires a pilot's presence at all times [2]. installations.
 The work presented in the paper is structured as follows. An topography, construction & infrastructure inspection, delivery
overview of the drone market is given in Section 2. Section 3 of goods, real-time monitoring of road traffic, providing
summarized the use of drones in the smart grid. The application wireless coverage, and extensive fauna census. The challenges
of drones for solar plant monitoring is given in Section 4. encountered for the operational use of drones are
Section 5 describes the role of the UAV in predictive and multidisciplinary. Thus, the drone system must be reliable,
application maintenance in the smart grid. The next revolution endowed with sufficient flight endurance, and a piloting mode
for the future of drones is given in Section 6. Section 7 provides in line with the shot's taking.
the conclusion of the work with possible future guidelines. Taking high-resolution images and videos using onboard
cameras to survey a given target area is a crucial challenge. This
II. DRONES ON ALL FRONTS helps to evaluate the magnitude of the damage to the
The drone's primary function is to extend human vision infrastructure. The shooting must be planned with a significant
beyond the natural horizon to perform at risk or in hostile overlap between images, without omitting the scene flown over.
environments. Drones earned an unsavory reputation. Most The drone vision system is often compared to human eyesight,
people associate them with expensive military aircraft. The but this machine vision isn't bound by biology and can be
military drones were the first to hit the skies, buzzing over programmed to see through obstacles and during the dark.
combat zones. Recent data shows that practical commercial For drones to be useful in many areas, one must use them in
applications will shape the future of drones. various contexts. In particular, when a drone fulfills a mission
The non-military drone industry should know a profound in an unfamiliar environment, the question of mobility is
change. The usage of drones in business operations has essential.
broadened in different sectors. More recently, civilian Many flight parameters will influence the result's quality,
applications have appeared, such as preventing forest fires, starting with the altitude and the distance between two flight
inspecting engineering structures, surveillance of motorway lines, conditioning the images' lateral overlap. Image
traffic, or collecting meteorological data. Depending on the processing, which will create three-dimensional models and
operational capabilities sought, the drones' size varies on a scale ortho-image mosaics, must meet the requirements in terms of
ranging from a few centimeters to several meters. precision and spectral quality required for the user-defined a
Simultaneously, their weight fluctuates between a few grams priori.
for the lightest and up to several tones for the heaviest. Modern photogrammetry is continuously evolving, and
Drone for recreational use and leisure activities should many questions remain about the configuration of these
continue to progress. However, a growing number of processing chains, depending on the objectives pursued. In
component suppliers worldwide are experiencing slowdowns or addition to their evolving environment, drones are
shutdowns in response to the COVID-19, affecting sub- characterized by their autonomy level. Again, each organism
components' availability in a snowball effect [9]. often uses the terminology of its own.
The global commercial drone market size is projected to  Evolution under "manual" control: when the UAV
increase. According to a new market research report titled, trajectory remotely piloted results at any time from
"Unmanned Aerial Vehicle (UAV) Market by Component commands sent by a remote pilot in real-time. The human
(Hardware, Software), Class (Mini UAVs, Micro UAVs), End- operator continuously controls the UAV system onboard,
User (Military, Commercial, Agriculture), type, capacity, and only from his direct observations. In this context, the UAV
Mode of Operation– Global Forecast to 2027", the UAV does not take any initiative and is dependent on the
market is expected to grow at a compound annual growth rate operator's orders.
(CAGR) of 14.1% from 2020 to 2027 to reach $21.8 billion by  Evolution "automatically": the flight is carried out without
2027 [9]. the intervention of a remote pilot.
In terms of volume, the market is expected to grow at a  Evolution "autonomously": The UAV fulfills the mission
CAGR of 16.2% from 2020 to 2027 to reach 13.2 million units assigned to it without human intervention, within a defined
by 2027. In that case, they should be largely overtaken by framework, while adapting to operational conditions and
professional models, estimated to them more than $5 billion. their environment.
And this strong trend should continue and gain momentum in In light of these different uses, the aerial drone has
the coming years, as the professional applications will develop. undeniable virtues. Unfortunately, unconscious or clumsy
A wide variety of types of drones exist, and their services tend to channel criticism on this socially and
characteristics define their fields of application. economically useful tool. It is, therefore, necessary to question
Several ambitious reforestation projects are underway the relevance of regulating these uses.
around the world. Depending on the objectives sought, different Some recreational drones require permission to fly. In all
drone systems have been developed, varying in size and capable cases, a certain number of rules condition their management.
of carrying a wide variety of sensors. The use of this technology Some areas are no-fly, and others subject to restrictions.
has, therefore, resulted in varying degrees of success. Compliance with the legislation in force, typically requiring the
Recent studies attempt to develop protocols for using drones drone platform's approval, the remote pilot's appropriate
in several applications such as archeology, search and rescue, training, and the flight authorization for each aerial mission, are
disaster management, remote sensing, precision agriculture, significant constraints on civilian drone deployments.
 III. DRONES TO INTERVENE IN THE ELECTRICITY NETWORK investment in renewable energy projects. Solar energy has been
The electricity distribution network's inspection is usually one of the most adopted among the different renewable energy
carried out by a series of remote mains tests, checks with solutions available. In the last decade (2009-2019), solar energy
binoculars from the ground, and lifting systems. Since the investments totaled 1.3 trillion dollars worldwide, representing
majority of lines are in hard-to-reach environments, current half of all assets in renewable energy.
methods are often expensive and time-consuming. Table 1. Smart grid infrastructure inspection and monitoring using
UAVs.
Easy to implement and robust design, the drones will
contribute to the security of electrical networks. Much of the UAV Type Application Payload/Altitude/End Sensor
current use of drones in the industry occurs in risky and urance Type
expensive activities, such as inspecting power transmission, MikroKopter Use UAVs for Up to 500 g /Up to 247 Laser
L4-ME vertical m/ 13-20 min. scanner
distribution lines, and power grids [6].
Quadcopter [4] inspection from
Working on a high-voltage tower is dangerous. The UAV- high rise
based solutions will eliminate the danger for electrical fitters, infrastructures
who do not have to climb the towers. such as street
lights
Due to its size and maneuverability, the drone makes it easy
Quadrotor, Inspection of the Less than 1 Kg/LAP/ Color
to monitor and inspect electrical equipment in hard-to-reach UAV [6] high voltage Less than 1 hour. and TIR
locations where UAVs can be carried in the back of an off-road power camera,
vehicle, four-wheel-drive vehicle, or snowmobile. This type of transmission GPS,
mission can also be performed by helicopters or planes, but the lines IMU
Fixed Wing Sketchy Less than 3 Kg /Up to HD
cost argument again proves the drone right without forgetting Aircraft, UAV inspection, 500 m/ Up to 50 min ultra-
the ecological benefit. Besides, automated UAVs allow [7] identifying the (50 km). wide
thousands of miles of facilities to be inspected without shutting defects of the angles
down operations while reducing the risk of accidents and power video
transmission camera
lowering the cost of their surveillance operations [7]. lines
By embedding an optical camera, thermal camera, and Quadrotor, Use cooperative Less than 6 kg /Up to TIR
sensors, drones make it possible to capture a great deal of data UAV [10] UAVs platform 200 m/ Up to 25 min camera,
and take high-resolution snapshots to carry out cartographies or for inspection (10 km). GPS
and diagnosis of
3D models - in complete safety. the power lines
Thanks to their GPS positioning and altitude-keeping system, infrastructures.
inspection is efficient and straightforward. Besides, drones can
operate in dusty areas and hot, cold, and windy weather, and A. The Growth of the Solar Energy Industry
they can work near a high voltage line. The drone is the ideal Economic factors have played an essential role in the
tool for power transmission and distribution lines and power increase in solar energy projects. The installation cost per kW
grids: has decreased by 73% in the last decade, from USD 4621 in
 Much faster than traditional line inspection. 2010 to USD 1210 in 2019 [11]. Solar photovoltaic (PV) plant
 It is sometimes impossible to access foot (remote, hilly, design involves implementing a series of features to achieve the
dense vegetation, etc.). lowest possible cost of electricity.
 The drone avoids having to resort to rope access Solar PV modules are made up of PV cells, most commonly
technicians systematically. Therefore, the safety gain is from silicon. Regular inspections of PV modules consist of
inevitable. inspecting each of the panels with a laser temperature gun. After
To simplify and speed up the process of inspecting these identifying solar panel failures, the operator should record these
power lines, the idea of using drones quickly came to the minds panels' locations for future repair. Due to the solar parks'
of companies in the energy sector. Given their ability to carry a dimensions, these traditional methods are inefficient and
very high-resolution camera while remaining mobile and detrimental to the operation and maintenance teams.
precise, drones prove to be perfect allies for this type of mission B. UAV-based Solar Inspections
[8]. Going underground - before the UAV thermography dawn, it
Table 1 summarizes some of the smart grid infrastructure was rarely done on more than 10% of a solar site inspection. In
inspections using UAVs. More specifically, this table presents a large solar farm, even 5 MW, more than 26,000 modules are
various kinds of UAVs used to inspect power transmission, required, and probably two to three technicians to complete the
distribution lines, and power grids, and the type of sensors work. Manual thermography is needed when the panel is under
deployed for each application, and the corresponding UAV the sun, and then the sun is behind the technician, thus casting
specifications in terms of payload, altitude, and endurance. a shadow. The drone does not suffer from that as it is far enough
away. The use of drones in the solar energy industry would
IV. DRONE FOR SOLAR PLANT MONITORING greatly benefit.
Due to growing concerns about climate change and carbon Drones equipped with thermal measurement technologies
emissions, many countries worldwide have increased their help the person in charge to obtain more accurate data and make
better decisions. in energy production. The faster the potential failures can be
One of the main reasons for integrating drones in the detected, the better the response and the lower the probability
examination of solar parks is reducing inspection time. Drone of significant system failures. Ultimately, a practical inspection
integration allows work teams to reduce inspection time by 70% in conjunction with a good maintenance plan results in better
compared to conventional methods. protection of the investment value and greater energy supply
With the ability to cover great distances and provide high- efficiency.
resolution images, drones have allowed companies from
different sectors to carry out safer and more effective V. TOWARDS PREDICTIVE AND APPLICATION MAINTENANCE
operations. Currently, many companies in the solar energy Beyond the inspection and surveillance that identify
sector use drones to improve their operations' efficiency and equipment requiring maintenance, drones can also act upstream
increase the quality of service offered to customers. Drones and downstream of this process.
increase the accuracy of solar inspections. Platforms such as the Upstream, thanks to their ability to collect data and carry out
DJI Matrice 210 RTK V2, equipped with thermal cameras such infrastructure maps at a very high level of detail. Using image
as the Zenmuse XT2, can inspect large areas of the solar park sensor-equipped device such as a digital camera, drones can
collecting RGB high-resolution thermal images (an example is power and enrich the digital twin that requires precise data.
shown in Fig. 1). The digital twin is a digital representation of a physical asset
that can simulate its operation and performance state.
Depending on the digital twin technologies, the simulations
predict different behavioral scenarios and make it possible to
anticipate risks and anomalies by notably estimating the
equipment's lifespan, thus ensuring maintenance predictive.
Drones are also called upon to act downstream of the process,
thus ensuring corrective maintenance. Dangerous tasks at
Fig. 1. (a) Quadcopter drone for monitoring solar plant; (b) Tri-copter
drone for monitoring solar plant (Source: Gregory Zeller, 2015
height, such as repairing or pruning along power lines, can be
http://www.innovateli.com/at-rotor-air-cams-hovering-over-the- left to drones. Enedis, a French manufacturer, launched a
future/) project to inspect, renovate, and maintain the electrical grids.
Several experiments have taken place: drones equipped with
C. Benefits of using Drones for Aerial Inspections
articulated arms have placed "avifauna" beacons (device
Energy inspections usually require thermal cameras, whereas protecting birds from the risk of collision with electrical
mapping operations require more precision in RTK modules. cables). Another type of drone was developed by Hélicéo
Aerial images provide a broad perspective on the solar park design. This automatic catamaran drone with dual technology
status and allow maintenance teams to focus on the points that of aerial or aquatic propulsion was used to layout the borehole
reflect potential failures. route and run an electric cable. Engineers can thus remotely and
Visual and thermal images provide the fault's precise precisely manipulate this drone to bury a power line under a
location, making fieldwork more accurate and less margin for river.
error. Thermal imaging makes it easy to detect defects based on
temperature anomalies between cells and panels. By combining A. Study case: Ice Detection on Electrical Power Cables
the thermal and visual RBG data, the operators can determine In northern countries, ice storms can cause major power
if these temperature anomalies result from physical factors such disruptions, such as the one that occurred on the North
as laminate, cracks, or dust; or if it is due to connection American ice storm of 1998 that left more than four million
problems such as inverters or wiring faults. The use of aerial customers in the dark with no electricity for periods varying
images contributes to the improvement in fault detection and from days to several weeks. In total, more than 20000 electricity
helps with the maintenance process. poles and 1000 steel towers collapsed under the weight of the
The information collected by the drones could be stored and ice. 3000 km of Hydro-Québec power lines were temporarily
organized for processing. Photogrammetry software out of use.
reconstructs thermal maps from the images obtained. Panel Detection of ice formation on power cables can help remove
placement is also adjusted using ground control points. After the ice before a significant problem occurs. The detection of
reconstruction and positioning adjustments, the resulting maps objects in images is an ancient theme of computer vision, which
are integrated into a Geographic Information System (GIS) has seen impressive progress since the 2000s, thanks to
where maintenance teams can quickly locate anomalies and approaches based on the learning of characteristic models of
rectify them before they have an appreciable effect on classes of objects. First, descriptors appeared (Haar wavelets,
production. set of points of interest, etc.), allowing the characterization of
Storing records of previous inspections is very helpful in regions of interest.
determining the causes of failures. In some cases, a solar cell Classification methods (AdaBoost, Random Forest, support
failure does not justify replacing the entire panel, but it is vector machine, etc.) are trained to detect objects in the areas
essential to record these cases to prevent significant losses. displaced in the request image's positions.
Early detection of faulty elements helps prevent inefficiencies Since 2010, these approaches have competed with methods
based on CNN, whose learning capacities are much more
significant due to having substantial learning bases and very
powerful computations.
Ice accretion on power cables poses severe risks to the
structure. To detect ice accumulation, we first tested many
approaches, taking advantage of existing methods, particularly
in OpenCV: segmentation by color, detection by gradients or
by contours, the combination of classifiers (AdaBoost, SVM)
applied to regions of interest or on pixels, etc. These various
methods did not achieve the target classification rates in these
types of applications, above 80%. However, the CNN
classifiers' significant difficulty is that there were no bases
available with images of ice accretions on the power
transmission line.
A CNN training then optimizes the network's coefficients
from a random initialization to minimize the classification error (a)
at the output. The two parts of the CNNs are trained
simultaneously. According to a gradient descent method, the
network's coefficients are modified to correct the classification
errors encountered. These gradients are backpropagated
through the network from the output layer. With a Region-
Based Convolutional Neural Networks (R-CNN) model, each
of the proposed areas of interest is received as input by the
CNN. Using a Faster R-CNN model, it is the starting image that
undergoes this convolution step. It is on the generated feature
maps that are obtained the proposals of regions likely to contain
the element targeted by the model. This is an improvement in
terms of runtime and power required.
In this work, for detection purposes, Faster-RCNN along (b)
with ResNet-101 architecture is used. In recent years, it has Fig. 3. Ice accretion detection results obtained by the proposed
Faster R-CNN, (a) the presence of the ice accumulation, (b) the
shown remarkable performance for various classification and
absence of the ice accumulation.
detection applications. It utilizes the same traditional network
for region proposal generation and detection. VI. PREDICTIONS FOR FUTURE TRENDS
Figure 2 illustrates one example of how Faster R-CNN
works. It is also known as a two-stage detection model as it So far, we have been able to have a fairly broad vision of the
essentially has two stages; the initial stage provides region state of the art of unmanned aerial vehicle technology. Still, its
anchors using Region Proposal Network (RPN). The second possibilities do not end here, indeed, as it is a technology that
stage is employed for electric power cable image classification evolves exponentially. The research on UAV-based smart grid
using bounding boxes and detected regions' information. monitoring is still in its infancy. We are in the first years of its
The simulation results for ice accretion detection are given in development. We think the next revolution will be substantial
Fig 3. for the future of drones. We point out several open issues as
further research directions.
 Drones and Augmented Reality: Airbus company, from its
DAR System (Drones and Augmented Reality for Aircraft
Inspections) department, is developing what aircraft safety will
be like in the coming years [12].
 Piloting drones with the mind: at the University of Florida,
the first race of mind-controlled drones has been held, in which
16 university students have participated to pilot the drones
through brain-computer interface technologies [13].
 Drones guided by Artificial Intelligence: A drone is not
autonomous, particularly beyond the line of sight, if it cannot
avoid the obstacles in its path and recalculate its route, or even
Fig. 2. The architecture of the proposed Faster R-CNN for ice accretion propose a new route, depending on the circumstances
detection. Two sub-networks take images as input and generate encountered. It must carry out its mission in the event of a
separate detections in classification confidence scores and bounding momentary loss of the satellite link with the ground station. The
box coordinates. intelligent drone is characterized by its ability to recognize
topographic data, integrate environmental data (winds, electrical damage evaluation.
temperature), and integrate information. Researchers from the A major technological challenge is to make drones more
University of Zurich, in Switzerland are developing new drones intelligent to maximize data collection and qualification during
to search and rescue lost mountaineers. their flight.
 The AI is useful in protecting the drone: several Research centers are working on drones' automation to
researchers are developing AI-based drones capable of improve their mission efficiency despite the difficulties
detecting certain malfunctions on automatic piloting. It can encountered, such as the loss of GPS position or an unforeseen
program the automatic return to base [14]. obstacle encounter. The drone must be able to become
 Hybrid drone development: this is the work carried out by autonomous in its decision-making. Another issue identified
the team of Wang Kangli and Ke Yijie, from the National relates to cybersecurity. Indeed, drones' connected world allows
University of Singapore, who have developed the U-Lion drone data collection, but it poses challenges in data protection.
as a hybrid model that can take off and land vertically like During its flight, the drone emits and receives various signals
helicopters and pass to a cruise flight like airplanes. The wings via wireless communication channels (5G, Wi-Fi, Bluetooth
can be fully folded or extended to promote type), which are vulnerable to hacking. This implies increased
 Drones capable of crossing the Pacific: this is the security to protect the transmitted data, which is critical in the
challenge that the Natilus company has set itself, developing energy sector.
freight transport drones that can autonomously cross the Pacific Despite these challenges mentioned, which are far from
Ocean. The first drone developed by the company is 9 meters exhaustive, drones' use is unanimous within the energy sector
long. Later the company plans to manufacture 24 drones and 42 due to the many benefits. These technologies will bring to
meters long, covering the route between Los Angeles and electric utility companies, such as infrastructure oversight,
Hawaii in about five hours and flying 6000 meters high. lower maintenance costs, and improving human decision-
 Drones that generate electricity: this is the project that making thanks to increasing data accuracy and reliability. All
Ampyx Power is developing. These are drones that are kept in that remains is to automate the piloting in complete safety!
the air, over the sea, tied to a rope from turbines, which they
pull when the wind pushes to generate electricity. REFERENCES
 Cardboard drones: this is one of the initiatives financed [1] S. Hayat, E. Yanmaz, and R. Muzaffar, Survey on unmanned aerial vehicle
networks for civil applications: A communications viewpoint, IEEE
by DARPA, the United States Defense Advanced Research Commun. Surveys Tuts., vol. 18, no. 4, pp. 2624_2661, 4th Quart., 2016.
Projects Agency, and which is developed by the company [2] H. Shakhatreh et al., Unmanned Aerial Vehicles (UAVs): A Survey on
Otherlab, to have a new system of disposable drones for the Civil Applications and Key Research Challenges, in IEEE Access, vol. 7,
pp. 48572-48634, 2019.
transport of merchandise to be used in high-risk areas or with [3] B. Li, Z. Fei, Y. Zhang and M. Guizani, Secure UAV Communication
limited access [15]. Networks over 5G, in IEEE Wireless Communications, vol. 26, no. 5, pp.
114-120, October 2019, doi: 10.1109/MWC.2019.1800458.
VII. CONCLUSION [4] I. Sa and P. Corke, Vertical infrastructure inspection using a quadcopter
and shared autonomy control, in Field and Service Robotics. Berlin,
Drones have gradually been adopted by the major players in Germany: Springer, pp. 219-23, 2014.
the energy sector, but several challenges remain to meet their [5] D. Jones, "Power line inspection, A UAV concept,'' in Proc. IEE Forum
Auton. Syst., pp. 1-8. 2005.
expectations. The business opportunity for using drones in the [6] L. F. Luque-Vega, B. Castillo-Toledo, A. Loukianov, and L. E. Gonzalez-
energy sector is impressive, although very few companies have Jimenez, Power line inspection via an unmanned aerial system based on the
joined in its implementation. quadrotor helicopter, in Proc. 17th MELE-CON, pp. 393-397, 2014.
[7] C. Sampedro, C. Martinez, A. Chauhan, and P. Campoy, "A supervised
More secure and less expensive than human intervention, approach to electric tower detection and classification for power line
drones can replace the inspection procedures traditionally inspection,'' in Proc. Int. Conf. Neural Netw, Jul. 2014, pp. 1970-1977.
performed by humans. [8] Z. Li, Y. Liu, R. Walker, R. Hayward, and J. Zhang, "Towards automatic
power line detection for a UAV surveillance system using pulse coupled
The drone allows the safe remote overflight of high voltage neural filter and an improved Hough transform,'' Mach. Vis. Appl., vol. 21,
power lines. Using drones means improved safety and flying no. 5, pp. 677-686, Aug. 2010.
continuously and quickly analyzing thousands of kilometers of [9] Metliculous, Unmanned Aerial Vehicle (UAV) Market by Component
(Hardware, Software), Class (Mini UAVs, Micro UAVs), End User
power cables. Until now, to inspect the power lines of an (Military, Commercial, Agriculture), Type (Fixed Wing, Rotary-Wing
industrial plant or the networks of high voltage lines, it was UAVs), Capacity, and Mode of Operation - Global Forecast to 2027", June
necessary to resort. 2020.
[10] C. Deng, S. Wang, Z. Huang, Z. Tan, and J. Liu, Unmanned aerial vehicles
Drones can restore high-quality and geotagged photos with a for power line inspection: A cooperative way in platforms and
powerful zoom, controllable from the ground station. They communications, J. Commun., vol. 9, no. 9, pp. 687-692, 2014.
carry out the visual inspection of cables and isolators' pylons. [11] S. Hosseini; M. Mesbahi, Energy-aware aerial surveillance for a long-
They allow electric utility companies to pay more attention to endurance solar-powered unmanned aerial vehicle, J. Guid., Control, Dyn.,
vol. 39, no. 9, pp. 1980-1993, 2016.
repairing identified issues rather than using valuable circuit [12] J. Pavlik, Drones, Augmented Reality and Virtual Reality Journalism:
downtime for troubleshooting. In addition to the safety and cost Mapping Their Role in Immersive News Content. Media and
benefits, the utility industry can now reap the benefits of 3D Communication, 8(3), 137-146, 2020.
[13] D. Tezza, et al. Brain-Controlled Drone Racing Game: A Qualitative
mapping by using professional UAVs to conduct surveys. Analysis. In: Fang X. (eds) HCI in Games. HCII. Lecture Notes in
Drones can also be used for post-disaster aerial assessment or Computer Science, vol 12211. Springer, Cham. 2020.
[14] D. Lee, W. Gyu La and H. Kim, Drone Detection and Identification System
using Artificial Intelligence, International Conference on Information and
Communication Technology Convergence, pp. 1131-1133, 2018.
[15] E. Ackerman, Swarms of Disposable Drones Will Make Critical Deliveries
and Then Vanish, IEEE Spectrum, Feb 2017.