Pith. sign in

REVIEW 4 major objections 5 minor 41 references

Dynamic Trajectory Adaptation for Efficient UAV Inspections of Wind Energy Units

T0 review · 4 major / 5 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read An automated vision-guided UAV trajectory planner for wind turbines claims 78% faster inspections, 17% shorter paths, 6% more blade coverage, and 68% less deviation from the optimal path.

desk verdict Clear enough pipeline description, but the headline 78/17/6/68% results are not backed by any real experiment and conflict with the paper's own Table I. read the letter →

arxiv 2411.17534 v1 pith:CREUUAJO submitted 2024-11-26 cs.RO cs.SYeess.SY

Serhii Svystun , Oleksandr Melnychenko , Pavlo Radiuk , Oleg Savenko , Anatoliy Sachenko , Andrii Lysyi This is my paper · ORCID
classification cs.ROcs.SYeess.SY
keywords unmannedaerialvehicleswindturbineinspectionautomatedtrajectoryplanningbladepitchangleimagesegmentationPIDcontrolcompensationopticalsensors
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper proposes a fully automated pipeline for planning UAV inspection flights around wind turbines and claims it outperforms manual piloting on every metric it reports. The core promise is that from a single image the system can segment the turbine into its components, measure each blade's pitch angle, and generate a flight trajectory that adapts to that angle in real time, with a PID controller correcting for wind and turbulence. According to the reported experiments, the method cuts total inspection time by 78%, shortens total trajectory length by 17%, raises average blade-surface coverage by 6%, and reduces average deviation from the optimal trajectory by 68%. If those numbers are correct, wind-farm inspections become faster, cheaper, and safer because they no longer require a human operator to plan and steer the flight.

What carries the argument

The load-bearing object is the dynamic trajectory function $S_i(t)$ in equation (8): a piecewise curve that takes the UAV from the initial point $S_{p,i}$ along a blade-adaptive path $f_{path}(S_{p,i}, t, \theta_i)$ for flight time $T_{fb}$ and then returns it to the start in $T_{cb}$. The blade angle $\theta_i$ is computed from the minimum-area rectangle around each segmented blade contour, and the PID law (equation (10)) produces per-coordinate corrections $u_x, u_y, u_z$ that keep the UAV on that path under wind. Static parts (tower, nacelle) reuse fixed precomputed trajectories; only the blades require dynamic adaptation. This combination is what the paper says produces the reported efficiency gains.

What would settle it

A concrete check is to run the full method on a single turbine in the field or in a high-fidelity simulator, log the flight, and recompute the four metrics. If the automated total inspection time is not about 22% of the manual time, or the mean deviation is not about 32% of the manual value, the central claim is contradicted. Simpler still: if the flight-execution block does not exist, Table I cannot be a measurement of the proposed method.

Watch

Extended reading notes

Core claim

The central claim is that a complete, inspection-ready flight trajectory can be derived automatically from the turbine's coordinates and a single segmented image, and that this automated trajectory beats manual UAV control. The paper's pipeline selects an initial viewing point opposite the drive mechanism, segments the image into blades, tower, and nacelle, filters the resulting contours, computes each blade's pitch angle from its minimum bounding rectangle, and then builds a blade-specific dynamic path that follows that angle while observing safe distance and optimal viewing angles. Static components are given fixed precomputed trajectories. The paper reports that across four scenarios covering different turbine counts, terrain, and wind speeds, the automated method produces a 78% reduction in inspection time, a 17% reduction in trajectory length, a 6% increase in blade surface coverage, and a 68% reduction in deviation from the optimal path.

Load-bearing premise

The load-bearing premise is that the numbers in Table I are measured outcomes of the proposed method, but the paper gives no experimental protocol and its own Section V lists the flight-execution block as future work; if Table I is estimated or planned, the four headline percentages are unsupported.

Editorial extensions

If this is right

  • If the results hold, wind-farm operators can run blade inspections without sending a pilot, which removes operator cost and the risk of pilot error from the data-collection step.
  • The reported 6% coverage increase means more blade area is imaged per pass, which should directly improve the recall of downstream defect-detection models.
  • PID-based wind compensation implies inspections can stay on a tight viewing plan even in moderate-to-strong wind, reducing repeat flights caused by off-path imagery.
  • The same pipeline, extended to multiple turbines, suggests fleet-level inspection schedules could be planned automatically from input coordinates alone.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A large share of the reported 78% time saving may come from parallelism: the first scenario compares three automated UAVs with one manual UAV, and the paper does not separate the effect of extra aircraft from the effect of better trajectories.
  • Because Section V treats flight execution (Block 2) as future work, the headline percentages are best read as projected outcomes of the planning pipeline, not as field measurements from the completed system.
  • The same pitch-angle-adaptive logic could transfer to inspection of other rotating structures, such as helicopter rotors or industrial fans, where an image-derived angle drives a dynamically updated flight path.
  • A straightforward extension would be a simulator study that replays the four scenarios with identical inputs and compares the proposed trajectories against the spiral and rectangular baselines cited in the paper; that would isolate how much of the gain is due to dynamic adaptation alone.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper proposes an automated UAV trajectory-determination method for wind turbine inspection, combining component segmentation (Detectron2/OpenCV), blade pitch-angle classification, trajectory generation with a PID controller, and static trajectories for tower/nacelle. The central claims are that the method reduces inspection time by 78%, trajectory length by 17%, increases blade surface coverage by 6%, and minimizes deviation from an optimal trajectory by 68%, based on computational experiments summarized in Table I.

Significance. If the reported improvements were substantiated by reproducible measurements, the method would be a practical contribution to automated wind-turbine inspection, which is an active and application-relevant area. The paper also proposes a reasonable high-level pipeline (segmentation, pitch classification, trajectory adaptation) and cites relevant related work. However, the quantitative contributions are not supported by the evidence presented: the numbers in Table I do not reproduce the claimed percentages, and the execution block needed to produce those measurements is explicitly stated to be future work. Without a validated experiment or a clear derivation, the headline results cannot be taken as established. The paper also provides no error analysis, no error bars, and no algorithmic details sufficient for replication.

major comments (4)
  1. [Section IV, Table I] The claimed aggregate percentages are not consistent with the values in Table I. Summing the four scenarios, manual inspection time totals 90+35+50+150 = 325 min versus 8+7+9+12 = 36 min for the automated method, a reduction of 88.9%, not 78%. Total trajectory length reduces from 5300 m to 4280 m, a 19.2% reduction, not 17%. Average blade coverage is 86.25% manually and 94.25% automatically, an increase of 8 percentage points, not 6%. Because the abstract, introduction, and conclusion all repeat the 78/17/6/68 figures, the headline quantitative claims are internally contradicted by the paper's own data.
  2. [Section V, Block 2] The paper states that 'The following research phase will focus on implementing Block 2, responsible for executing the flight around the WEU along the specified trajectories.' Block 2 is precisely the component that would generate the measured inspection times, coverage values, and trajectory deviations reported in Table I. No simulator, dataset, or experimental protocol is described anywhere in the manuscript. Consequently, Table I cannot be regarded as a measurement of the proposed method, and the reported performance improvements are not supported by the presented evidence.
  3. [Section III, Eqs. (8), (10), and Step 1.5] The trajectory-generation equations are under-specified. In Eq. (8), the functions f_path and f_cm are asserted without definitions of how the trajectory is shaped by blade geometry, safe distances, or viewing angles; the notation S_p,i and T_fb/T_cb is not sufficient to reconstruct an implementable algorithm. Similarly, Eq. (10) gives generic PID correction integrals without specifying the error variables e_x, e_y, e_z, the controller gains, the UAV dynamic model, or how the corrections are combined with the nominal trajectory. These omissions prevent replication and independent verification of the claimed accuracy and deviation reductions.
  4. [Table I, Scenario 1] The first scenario compares a manual method using 1 UAV with the proposed automated method using 3 UAVs. This confounds the comparison: the time reduction could be due to the additional UAVs rather than to the method being evaluated. The other scenarios do not state the number of UAVs for the automated method, so the comparison basis across the table is not consistent. This further undermines the validity of the aggregate improvement percentages.
minor comments (5)
  1. [Section III, Eq. (7)] The blade pitch angle is defined as the arctangent of the line through the top and bottom points of the minimum bounding rectangle, but the classification thresholds in Step 1.4 assume angles in the range [0°, 180°] without discussing how the arctangent result is wrapped to that range or how the rectangle's orientation is unambiguously chosen. This may produce inconsistent classifications for blades with similar inclinations.
  2. [Fig. 6] The axis labels and legend in Fig. 6 are garbled in the provided text (e.g., 'Traditional roposed' and unreadable axis values), making the figure difficult to interpret; a clean vector version with proper labels should be provided.
  3. [References] Reference [11] is cited as previous work by the authors, but the citation details at the end of the paper list the title 'Intelligent integrated system for fruit detection using multi-UAV imaging and deep learning,' which is unrelated to wind turbine inspection; the in-text comparison to 'similar improvements in inspection efficiency' is unclear and should be clarified.
  4. [Section IV, Fig. 6 caption] The caption states that the figure compares 'traditional (blue) and proposed (green)' methods but the plotted data for coverage and deviation appear to be computed from Table I; no error bars or statistical replicates are indicated, so the figure does not convey the uncertainty of the measurements.
  5. [Section I, Contributions] The introduction lists the 78/17/6/68 percent claims as contributions before any evidence is presented, but the body text and Table I do not support these exact numbers; the claims should be stated only after presenting consistent experimental results.

Circularity Check

0 steps flagged · score 0.0 of 10

No circular derivation found: the claimed percentages are asserted from Table I and are internally inconsistent with it, while Section V states flight execution is future work; these are validity and reproducibility defects, not circularity.

full rationale

The claimed 78%/17%/6%/68% improvements are presented as read off Table I, but Table I's aggregate values imply 88.9% time reduction, 19.2% trajectory-length reduction, an 8-percentage-point coverage increase, and a 69.3% deviation reduction, and Section V states that Block 2, the flight-execution block, is future work. These are serious evidentiary and internal-consistency problems, but they are not circularity: no equation in the method is defined in terms of the reported performance metrics, no parameter is fitted to a subset of the data and then renamed a prediction, and the one self-citation (reference [11], 'Similar improvements in inspection efficiency have been demonstrated by our previous work') is a rhetorical comparison rather than a load-bearing premise that forces the stated outcomes. The trajectory-planning equations (1)-(10) and the Detectron2/OpenCV pipeline rely on external tools and stated geometric definitions; there is no exhibited reduction by which the reported percentages are equivalent to the method's inputs by construction. Therefore no specific circular step can be quoted, and the appropriate finding is no significant circularity.

Assumptions & free parameters 3 free parameters · 6 assumptions · 0 invented entities

The proposed method depends on several unstated parameters and strong domain assumptions about perception, 3D angle estimation, and controller behavior. The experimental evidence is an unexplained comparison table without code, data, or protocol.

free parameters (3)
  • Contour area threshold A_threshold
    Used in Eq. (6) to filter contours; no value, tuning procedure, or sensitivity analysis is given.
  • PID gains K_P, K_I, K_D
    Used in Eq. (10) to compensate wind; no values or tuning method are given, yet the 68% deviation reduction is attributed to them.
  • Blade pitch angle classification boundaries (30/60/90/120/150 degrees) = 30, 60, 90, 120, 150
    Chosen by hand in Step 1.4 with no justification; they determine trajectory switch behavior.
assumptions (6)
  • domain assumption A wind turbine inspection zone can be represented as a sphere with center C_i and radius R_i.
    Eq. (1)-(3); assumes the whole inspection space is a sphere around blade tips, which ignores tower/nacelle geometry and obstacles.
  • domain assumption Detectron2 segmentation gives accurate pixel masks for blades, tower, and nacelle from a single image.
    Step 1.2; off-the-shelf model, no fine-tuning data or accuracy metrics provided.
  • domain assumption The blade pitch angle computed from a 2D minimum bounding rectangle (Eq. 7) corresponds to the true 3D blade orientation.
    Step 1.4; a 2D image angle depends on camera pose and perspective and is not a 3D orientation without extrinsic calibration.
  • ad hoc to paper The trajectory functions f_p and f_cm in Eq. (8) produce collision-free, flyable paths and safe distances.
    Eq. (8) is a generic interpolation formula; no flight dynamics, obstacle constraints, or safety distances are specified.
  • domain assumption A PID controller with the stated form can compensate for wind and turbulence during inspection.
    Eq. (10); no controller stability analysis, gain values, or wind disturbance model is provided, and flight execution is stated to be future work.
  • ad hoc to paper Table I compares equivalent inspection conditions between manual and automated methods.
    Table I; Scenario 1 compares 1 manual UAV against 3 automated UAVs, so the conditions are not equivalent.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Dynamic Trajectory Adaptation for Efficient UAV Inspections of Wind Energy Units." pith.science (2026). https://pith.science/paper/CREUUAJO

@misc{pith2026241117534,
  author       = {Pith},
  title        = {Pith review of: Dynamic Trajectory Adaptation for Efficient UAV Inspections of Wind Energy Units},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/CREUUAJO}},
  note         = {Machine review of arXiv:2411.17534}
}
read the original abstract

The research presents an automated method for determining the trajectory of an unmanned aerial vehicle (UAV) for wind turbine inspection. The proposed method enables efficient data collection from multiple wind installations using UAV optical sensors, considering the spatial positioning of blades and other components of the wind energy installation. It includes component segmentation of the wind energy unit (WEU), determination of the blade pitch angle, and generation of optimal flight trajectories, considering safe distances and optimal viewing angles. The results of computational experiments have demonstrated the advantage of the proposed method in monitoring WEU, achieving a 78% reduction in inspection time, a 17% decrease in total trajectory length, and a 6% increase in average blade surface coverage compared to traditional methods. Furthermore, the process minimizes the average deviation from the optimal trajectory by 68%, indicating its high accuracy and ability to compensate for external influences.

Figures

Figures reproduced from arXiv: 2411.17534 by the authors.

Figure 1
Figure 1. The scheme of the proposed method reflects the UAV-based wind turbine inspection process, from input data of spatial coordinates through flight path determination and execution, resulting in video frames. Block 1. Determination of flight trajectories around wind turbine components. Step 1.1. Determination of the initial flight point. The efficiency of data collection from WEUs using UAVs significantly depends on the… view at source ↗
Figure 2
Figure 2. The spatial positioning and trajectory planning for a UAV inspecting a wind turbine with two perspectives: (a) a frontal view with the UAV determining its initial flight point 𝑆𝑝 within the inspection zone 𝑍𝑖 , and (b) a side view emphasizing the calculated flight trajectory around the turbine blades, highlighting the optimized path 𝐶𝑖 for component coverage. where n represents the number of blades, and ( x y z k k … view at source ↗
Figure 3
Figure 3. The sequential stages of image processing in the proposed method: (a) the original captured image, (b) the segmented image highlighting individual components, (c) the transformed image isolating relevant features, and (d) the final contoured image used for detailed analysis and trajectory planning [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: The classification of wind turbine blade tilt types based on their angular orientation in the x-y plane: (a) acute tilt with angles ranging from 30° to 60° and from 120° to 150°, (b) horizontal tilt with angles from 0° to 30° and from 150° to 180°, and (c) vertical til…
Figure 5
Figure 5. Figure 5: This figure details the proposed method for UAV-based wind turbine inspection, highlighting the stages from assigning turbine coordinates, component segmentation, and contour detection to determining flight trajectory and blade orientation. As shown in [PITH_FULL_IMAG…
Figure 6
Figure 6. Figure 6: This figure compares traditional (blue) and proposed (green) UAV-based wind turbine inspection methods across four scenarios, highlighting total inspection time (a), trajectory length (b), surface coverage (c), and deviation from optimal trajectory (d). The proposed me…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 21 canonical work pages

  1. [1]

    4a): (   )30 ,60 120 ,150 .i      

    Acute tilt (Fig. 4a): (   )30 ,60 120 ,150 .i      

    work page

  2. [2]

    4b):    0 ,30 150 ,180 .i    

    Horizontal tilt (Fig. 4b):    0 ,30 150 ,180 .i    

    work page

  3. [3]

    Vertical tilt (Fig. 4c): ( )60 ,120 .i    The output data of this section consists of the set of values determined for each individual blade, which we represent as   1 2 3,,i i i i   = for each i-th blade. Fig. 3. The sequential stages of image processing in the proposed method: (a) the original captured image, (b) the segmented image highlighti...

    work page 1900

  4. [4]

    Fault diagnosis methods based on machine learning and its applications for wind turbines: A review,

    T. Sun, G. Yu, M. Gao, L. Zhao, C. Bai, and W. Yang, “Fault diagnosis methods based on machine learning and its applications for wind turbines: A review,” IEEE Access, vol. 9, pp. 147481 –147511, 2021, doi: 10.1109/ACCESS.2021.3124025. Traditional roposed Inspection Time min 0 20 00 0 0 0 20 0 Scenario Scenario 2 Scenario 3 Scenario Total Inspection Time ...

    arXiv 2021

  5. [5]

    Land-based wind market report: 2023 Edition

    R. Wiser et al., “Land-based wind market report: 2023 Edition”, Office of Scientific and Technical Information (OSTI), Oak Ridge, 1996790, Aug. 2023. Accessed: Aug. 25, 2024. [Online]. Available: https://doi.org/10.2172/1996790

    work page doi:10.2172/1996790 2023

  6. [6]

    Balancing profitability of energy production, societal impacts and biodiversity in offshore wind farm design,

    E. A. Virtanen et al., “Balancing profitability of energy production, societal impacts and biodiversity in offshore wind farm design,” Renewable Sustain. Energy Rev., vol. 158, p. 112087, Apr. 2022, doi: 10.1016/j.rser.2022.112087

    work page arXiv 2022

  7. [7]

    Diagnosis of wind turbine faults with transfer learning algorithms,

    W. Chen, Y. Qiu, Y. Feng, Y. Li, and A. Kusiak, “Diagnosis of wind turbine faults with transfer learning algorithms,” Renew. Energy, vol. 163, pp. 2053–2067, Jan. 2021, doi: 10.1016/j.renene.2020.10.121

    work page doi:10.1016/j.renene.2020.10.121 2021

  8. [8]

    Approach to face recognition using neural networks,

    I. aliy, A. Sachenko, V. Koval, and Y. Kurylyak, “Approach to face recognition using neural networks,” in 2005 IEEE Int. Conf. Intell. Data Acquisition Adv. Comput. Syst.: Technol. Appl. (IDAACS), Sofia, Bulgaria, Sep. 5–7, 2005. New York, NY, USA: IEEE, 2005, pp. 112-

    work page 2005

Show all 41 references
  1. [9]

    Wind turbine condition monitoring: technical and commercial challenges,

    W. Yang, . J. Tavner, C. J. Crabtree, Y. Feng, and Y. Qiu, “Wind turbine condition monitoring: technical and commercial challenges,” Wind Energy, vol. 17, no. 5, pp. 673–693, 2014, doi: 10.1002/we.1508

    2014 doi

  2. [10]

    UAV and IoT -based systems for the monitoring of industrial facilities using digital twins: Methodology, reliability models, and application,

    Y. Sun et al. , “UAV and IoT -based systems for the monitoring of industrial facilities using digital twins: Methodology, reliability models, and application,” Sensors, vol. 22, no. 17, Art. no. 17, 2022, doi: 10.3390/s22176444

    2022 doi

  3. [11]

    Recent technology and challenges of wind energy generation: A review,

    S. Roga, S. Bardhan, Y. Kumar, and S. K. Dubey, “Recent technology and challenges of wind energy generation: A review,” Sustain. Energy Technol. Assessments , vol. 52, no. C, p. 102239, Aug. 2022, doi: 10.1016/j.seta.2022.102239

    2022

  4. [12]

    An internet of drone -based multi -version post -severe accident monitoring system: Structures and reliability,

    H. Fesenko, V. Kharchenko, A. Sachenko, R. Hiromoto, and V. Kochan, “An internet of drone -based multi -version post -severe accident monitoring system: Structures and reliability,” in Dependable IoT for Human and Industry, 1st ed., in Computer Science, Engineering & Technolog...

    2018 doi

  5. [13]

    UAV assistance paradigm: State -of-the-art in applications and challenges,

    B. Alzahrani, O. S. Oubbati, A. Barnawi, M. Atiquzzaman, and D. Alghazzawi, “UAV assistance paradigm: State -of-the-art in applications and challenges,” J. Netw. Comput. Appl. , vol. 166, p. 102706, Sep. 2020, doi: 10.1016/j.jnca.2020.102706

    2020

  6. [14]

    Towards accurate image stitching for drone -based wind turbine blade inspection,

    C. Yang, X. Liu, H. Zhou, Y. Ke, and J. See, “Towards accurate image stitching for drone -based wind turbine blade inspection,” Renew. Energy, vol. 203, pp. 267 –279, Feb. 2023, doi: 10.1016/j.renene.2022.12.063

    2023 doi

  7. [15]

    Unmanned aerial vehicle UAV -assisted damage detection of wind turbine blades: A review,

    Z. Zhang and Z. Shu, “Unmanned aerial vehicle UAV -assisted damage detection of wind turbine blades: A review,” Energies, vol. 17, no. 15, Art. no. 15, Jan. 2024, doi: 10.3390/en17153731

    2024 doi

  8. [16]

    Intelligent integrated system for fruit detection using multi -UAV imaging and deep learning,

    O. Melnychenko, L. Scislo, O. Savenko, A. Sachenko, and P. Radiuk, “Intelligent integrated system for fruit detection using multi -UAV imaging and deep learning,” Sensors, vol. 24, no. 6, p. 1913, Mar. 2024, doi: 10.3390/s24061913

    1913 doi

  9. [17]

    A systematic literature review SLR on autonomous path planning of unmanned aerial vehicles,

    A. ul Husnain, N. Mokhtar, N. Mohamed Shah, M. Dahari, and M. Iwahashi, “A systematic literature review SLR on autonomous path planning of unmanned aerial vehicles,” Drones, vol. 7, no. 2, Art. no. 2, Feb. 2023, doi: 10.3390/drones7020118

    2023 doi

  10. [18]

    From ID to swarms: A decade of advancements in drone control and path planning - A systematic review (2013–2023 ,

    B. Cetinsaya, D. Reiners, and C. Cruz-Neira, “From ID to swarms: A decade of advancements in drone control and path planning - A systematic review (2013–2023 ,” Swarm Evol. Computation, vol. 89, p. 101626, Aug. 2024, doi: 10.1016/j.swevo.2024.101626

    2013

  11. [19]

    Apple detection with occlusions using modified YOLOv5-v ,

    O. Melnychenko, O. Savenko, and . Radiuk, “Apple detection with occlusions using modified YOLOv5-v ,” in 2023 IEEE 12th Int. Conf. Intell. Data Acquisition Adv. Comput. Syst.: Technol. Appl. (IDAACS), Dortmund, Germany, Sep. 7 –9, 2023. New York, NY, USA: IEEE, 2023, pp. 107–1...

    2023

  12. [20]

    Information support of the remote nitrogen monitoring system in agricultural crops,

    V. Lysenko, O. Opryshko, D. Komarchuk, N. Pasichnyk, N. Zaets, and A. Dudnyk, “Information support of the remote nitrogen monitoring system in agricultural crops,” Int. J. Comput., vol. 17, no. 1, pp. 47–54, Mar. 2018, doi: 10.47839/ijc.17.1.948

    2018 doi

  13. [21]

    Automatic defect detection in wind turbine blade images: Model benchmarks and re-annotations,

    I. Gohar, J. See, A. Halimi, and W. K. Yew, “Automatic defect detection in wind turbine blade images: Model benchmarks and re-annotations,” in 2023 IEEE Int. Conf. Multimedia Expo Workshops (ICMEW), Brisbane, Australia, Jul. 10–14, 2023. New York, NY, USA: IEEE Inc., 2023, pp....

    2023

  14. [22]

    Distributed trajectory planning for a formation of aerial vehicles inspecting wind turbines,

    D. érez, A. Alcántara, and J. Capitán, “Distributed trajectory planning for a formation of aerial vehicles inspecting wind turbines,” in 2022 Int. Conf. Unmanned Aircr. Syst. (ICUAS) , Dubrovnik, Croatia, Jun. 21–24, 2022. New York, NY, USA: IEEE, 2022, pp. 646 –654. doi: 10.1...

    2022

  15. [23]

    Image based surface damage detection of renewable energy installations using a unified deep learning approach,

    A. Shihavuddin et al. , “Image based surface damage detection of renewable energy installations using a unified deep learning approach,” Energy Rep. , vol. 7, pp. 4566 –4576, Nov. 2021, doi: 10.1016/j.egyr.2021.07.045

    2021 doi

  16. [24]

    Inspection of floating offshore wind turbines using multi -rotor unmanned aerial vehicles: Literature review and trends,

    K. Zhang, V. akrashi, J. Murphy, and G. Hao, “Inspection of floating offshore wind turbines using multi -rotor unmanned aerial vehicles: Literature review and trends,” Sensors, vol. 24, no. 3, Art. no. 3, Jan. 2024, doi: 10.3390/s24030911

    2024 doi

  17. [25]

    Review on the advancements in wind turbine blade inspection: integrating drone and deep learning technologies for enhanced defect detection,

    M. Memari, P. Shakya, M. Shekaramiz, A. C. Seibi, and M. A. S. Masoum, “Review on the advancements in wind turbine blade inspection: integrating drone and deep learning technologies for enhanced defect detection,” IEEE Access, vol. 12, pp. 33236 –33282, 2024, doi: 10.1109/ACCE...

    2024

  18. [26]

    Autonomous wind-turbine blade inspection using LiDAR -equipped unmanned aerial vehicle,

    M. Car, L. Markovic, A. Ivanovic, M. Orsag, and S. Bogdan, “Autonomous wind-turbine blade inspection using LiDAR -equipped unmanned aerial vehicle,” IEEE Access, vol. 8, pp. 131380 –131387, 2020, doi: 10.1109/ACCESS.2020.3009738

    2020

  19. [27]

    Understanding the human factors challenge of handover between levels of automation for uncrewed air systems: a systematic literature review,

    B. Grindley, K. J. Parnell, T. Cherett, J. Scanlan, and K. L. Plant, “Understanding the human factors challenge of handover between levels of automation for uncrewed air systems: a systematic literature review,” Transp. Planning Technol. , pp. 1 –26, 2024, doi: 10.1080/0308106...

    2024

  20. [28]

    ath planning techniques for unmanned aerial vehicles: A review, solutions, and challenges,

    S. Aggarwal and N. Kumar, “ ath planning techniques for unmanned aerial vehicles: A review, solutions, and challenges,” Comput. Commun., vol. 149, pp. 270 –299, Jan. 2020, doi: 10.1016/j.comcom.2019.10.014

    2020 doi

  21. [29]

    Data fusion and ensemble learning for advanced anomaly detection using multi-spectral RGB and thermal imaging of small wind turbine blades,

    M. Memari, M. Shekaramiz, M. A. S. Masoum, and A. C. Seibi, “Data fusion and ensemble learning for advanced anomaly detection using multi-spectral RGB and thermal imaging of small wind turbine blades,” Energies, vol. 17, no. 3, Art. no. 3, Jan. 2024, doi: 10.3390/en17030673

    2024 doi

  22. [30]

    Slice -aided defect detection in ultra high -resolution wind turbine blade images,

    I. Gohar, A. Halimi, J. See, W. K. Yew, and C. Yang, “Slice -aided defect detection in ultra high -resolution wind turbine blade images,” Machines, vol. 11, no. 10, Art. no. 10, Oct. 2023, doi: 10.3390/machines11100953

    2023 doi

  23. [31]

    A self-organized automated system to control unmanned aerial vehicles for object detection,

    O. Melnychenko and O. Savenko, “A self-organized automated system to control unmanned aerial vehicles for object detection,” in 4th Int. Work. on Intell. Inform. Technol. & Syst. of Inform. Secur. (IntelITSIS), Khmelnytskyi, Ukraine, Mar. 22–24, 2023. Aachen, Germany: CEUR- WS...

    2023

  24. [32]

    facebookresearch/detectron2. (Jul. 17, 2024). Python. Meta Research. Accessed: Aug. 21, 2024. [Online]. Available: https://github.com/facebookresearch/detectron2

    2024

  25. [33]

    Canny -based method of image contour segmentation,

    D. Zahorodnia, Y. igovsky, and . Bykovyy, “Canny -based method of image contour segmentation,” Int. J. Comput. , vol. 15, no. 3, pp. 200–205, Sep. 2016, doi: 10.47839/ijc.15.3.853

    2016 doi

  26. [34]

    LiDAR -based automated UAV inspection of wind turbine rotor blades,

    C. Castelar Wembers, J. Pflughaupt, L. Moshagen, M. Kurenkov, T. Lewejohann, and G. Schildbach, “LiDAR -based automated UAV inspection of wind turbine rotor blades,” J. Field Robot., vol. 41, no. 4, pp. 1116–1132, 2024, doi: 10.1002/rob.22309

    2024 doi

  27. [35]

    from the OpenCV library. As a result, we obtain a set of contours mC : 1 , m mi i CC = = () where iC is the set of contours identified in the mask , .biM To remove noise and irrelevant contours from mC calculated by (5), an area-based filter is applied: ( )filtered threshold ...

  28. [36]

    Research on automatic path planning of wind turbines inspection based on combined UAV,

    Z. Li, J. Wu, J. Xiong, and B. Liu, “Research on automatic path planning of wind turbines inspection based on combined UAV,” in 2024 IEEE Int. Symp. Broadband Multimedia Syst. Broadcast. (BMSB), Toronto, ON, Canada, Jun. 19–21, 2024. New York, NY, USA: IEEE, Jun. 2024, pp. 1–6...

    2024

  29. [37]

    UAV trajectory planning via viewpoint resampling for autonomous remote inspection of industrial facilities,

    H. Liu, Y. . Tsang, C. K. M. Lee, and C. H. Wu, “UAV trajectory planning via viewpoint resampling for autonomous remote inspection of industrial facilities,” EEE Trans. Ind. Inform. , vol. 20, no. 5, pp. 7492–7501, May 2024, doi: 10.1109/TII.2024.3361019

    2024

  30. [39]

    OpenCV with ython,

    S. Gollapudi, “OpenCV with ython,” in Learn Computer Vision Using OpenCV, S. Gollapudi, Ed., Berkeley, CA: Apress, 2019, pp. 31 –50. doi: 10.1007/978-1-4842-4261-2_2

    2019 doi

  31. [40]

    Developing a faster pattern matching algorithms for intrusion detection system,

    I. Obeidat and M. AlZubi, “Developing a faster pattern matching algorithms for intrusion detection system,” Int. J. Comput., vol. 18, no. 3, pp. 278–284, Sep. 2019, doi: 10.47839/ijc.18.3.1520

    2019 doi

  32. [41]

    ID control of quadrotor UAVs: A survey,

    I. Lopez -Sanchez and J. Moreno -Valenzuela, “ ID control of quadrotor UAVs: A survey,” Annu. Rev. Control, vol. 56, p. 100900, Jan. 2023, doi: 10.1016/j.arcontrol.2023.100900

    2023

  33. [115]

    doi: 10.1109/IDAACS.2005.282951

    2005

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.