Pith. sign in

REVIEW 5 major objections 5 minor 42 references

HOTSPOT-YOLO: A Lightweight Deep Learning Attention-Driven Model for Detecting Thermal Anomalies in Drone-Based Solar Photovoltaic Inspections

T0 review · 5 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

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

Pith's one-line read The paper claims that HOTSPOT-YOLO—a YOLOv11 variant with an EfficientNet backbone and Squeeze-and-Excitation attention—detects thermal anomalies in drone-based PV images at 90.8% mean average precision, beating YOLOv11m by 5.9 points with

desk verdict A plausible architecture and some nice qualitative demos, but the central mAP claim is unverifiable because the benchmark protocol is missing and the tables contradict the text; not ready for review. read the letter →

arxiv 2508.18912 v1 pith:P4JQNDGZ submitted 2025-08-26 eess.IV cs.AIcs.LG

Mahmoud Dhimish This is my paper · ORCID
classification eess.IVcs.AIcs.LG
keywords thermalanomalydetectionsolarphotovoltaicinspectionsHOTSPOT-YOLOYOLOv11EfficientNetSqueeze-and-Excitationattentiondronethermographymeanaverageprecision
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

Solar panels develop small hotspots that drone thermal cameras can catch only if an automated detector is both accurate enough to see subtle temperature differences and light enough to run in real time. HOTSPOT-YOLO is the paper's attempt to build that detector by grafting an EfficientNet backbone and Squeeze-and-Excitation attention onto YOLOv11. The paper reports 90.8% mean average precision on thermal PV images, a 5.9-point gain over YOLOv11m, with 2.12 million fewer parameters and real-time inference. The practical stake is straightforward: if the result holds, routine drone sweeps of utility-scale solar plants could flag faulty modules sooner and run on cheaper onboard hardware.

What carries the argument

The load-bearing mechanism is channel recalibration: Squeeze-and-Excitation blocks compress each feature map to a per-channel scalar via global average pooling, learn two fully connected layers that assign importance weights, and multiply those weights back into the feature maps. That is what lets the model emphasize thermally significant channels while suppressing background. Around this core, EfficientNet's depthwise separable convolutions keep the parameter and FLOP counts low, and the feature-aggregation block adds low-, mid-, and high-level feature maps so small hotspots and larger defective regions are seen by the same detection head.

What would settle it

Train all six comparison detectors and HOTSPOT-YOLO on the same public thermal PV image set with identical splits, augmentation, optimizer settings, epochs, and random seeds. If YOLOv11m reaches or exceeds 90.8% mAP, or if HOTSPOT-YOLO falls to the 84.9% level, the claimed architectural advantage collapses and the difference would instead reflect protocol differences.

Watch

Extended reading notes

Core claim

On its own terms, the paper advances one architectural recipe: keep the YOLOv11 detection head, replace feature extraction with depthwise-separable EfficientNet blocks, interleave SE blocks that recalibrate feature channels after global average pooling, and fuse low-, mid-, and high-level feature maps before prediction. Trained for 200 epochs with Adam and cosine learning-rate decay on a public collection of thermal PV images, the model is reported to reach 90.8% mAP@0.5, beat YOLOv11m by 5.9 points, and run at 25.22 ms per image with 36.10M parameters and 25.53 GFLOPs. The paper also reports that confidence falls under heavy blur and for ground-based oblique images, which it explains as a c

Load-bearing premise

The benchmark comparison assumes every baseline was trained and evaluated under exactly the same data splits, preprocessing, augmentation, and hyperparameters as HOTSPOT-YOLO, but the paper specifies training details only for HOTSPOT-YOLO.

Editorial extensions

If this is right

  • At 90.8% mAP, automated aerial inspection would catch a larger share of small hotspot defects than the 84.9% YOLOv11m baseline, reducing missed faults during routine surveys.
  • At 36.10M parameters and 25.53 GFLOPs, the model fits the compute envelope of drone-class hardware while still running at 25.22 ms per frame.
  • The reported stability under brightness, contrast, grayscale, and blur transformations means a single deployed model could handle variable weather and camera settings without per-site retraining.
  • The weaker results on ground-based oblique images imply the model works best for drone flight paths as trained, and that expansion to handheld inspections would need additional fine-tuning data.

Reading between the lines

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

  • A natural next step the paper does not report is a controlled re-run of all baselines under the exact same training protocol; that experiment alone would separate the architecture's contribution from setup differences.
  • The EfficientNet-plus-SE recipe is not solar-specific; it should transfer to other small-anomaly thermal domains, and can be tested directly on public thermal fault datasets for power equipment or buildings.
  • Appendix A's center-weighted bounding-box distribution suggests a per-location stratified mAP would reveal whether accuracy holds across the full drone frame.
  • Since blur clearly lowers confidence, blur augmentation during training is a direct, testable fix that might close that gap.
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

5 major / 5 minor

Summary. The paper proposes HOTSPOT-YOLO, a YOLOv11 variant that replaces the backbone with EfficientNet and adds Squeeze-and-Excitation attention, targeting thermal anomaly detection in drone-based PV inspection. The authors report 90.8% mAP@0.5 on a dataset assembled from Roboflow, a 5.9-point improvement over YOLOv11m, and lower parameter/FLOP counts. The manuscript includes qualitative robustness tests under brightness/contrast, grayscale, blur, and ground-based imaging, and a benchmark against six detectors in Table 3. The central claim is that the proposed architecture is both more accurate and lighter than the compared baselines.

Significance. If the reported results were reproducible, the contribution would be practically useful: a lightweight, real-time thermal anomaly detector for PV inspections. The qualitative robustness demonstrations in Figures 7-9 are a genuine strength, and the SE/multi-scale aggregation formulation is clearly described. However, the paper's central numerical claims cannot be independently verified: the dataset is unnamed, code and weights are absent, baseline training protocols are not documented, and several internal inconsistencies affect the architecture and split sizes. No ablations support the attribution of the gains to the EfficientNet backbone and SE attention. As an empirical benchmark paper, these issues are load-bearing rather than cosmetic.

major comments (5)
  1. [§2.3 and Table 2] The dataset is not identifiable and the split sizes are contradictory. §2.3 says 'publicly available thermal imagery datasets sourced from Roboflow' but names no dataset, version, license, or class balance. It also states 6000 training / 720 validation / 450 test images, while Table 2 lists 3600 training and 470 validation images. Since every mAP number in the paper depends on this dataset, the discrepancy must be corrected and the dataset must be identified or released. No code or trained weights are provided either.
  2. [§3.6, Table 3] The baseline benchmark protocol is undocumented. Table 3 reports mAP, parameters, FLOPs, and inference time for Cascade R-CNN, CenterNet, Faster R-CNN, YOLOv5m, YOLOv9m, and YOLOv11m, but the manuscript never states how these models were trained, on which split, with which hyperparameters, augmentations, data preprocessing, or seeds. Only HOTSPOT-YOLO's settings appear in Table 2. Without this information, the claimed 5.9-point improvement over YOLOv11m cannot be attributed to the architecture; it may reflect protocol differences rather than the proposed modifications.
  3. [Table 1] The layer-wise architecture table is internally inconsistent. The first Conv2D layer (3x3, 3→32 channels) has 864 weights plus bias, i.e., ~0.0009M parameters, not the listed 0.9M. The depthwise/pointwise counts are similarly off by roughly 1000x, and the entries do not sum to the reported 36.10M. The preprocessing row also lists 224x224x3 output, contradicting the 640x640 input used throughout the paper. Consequently the claimed parameter reduction and 'lightweight' characterization are not supported by the provided specification.
  4. [§2.3 and §3.6] The headline result is reported on the validation split, not a held-out test set: §2.3 says the final mAP of 90.8% was obtained 'using the validation dataset.' Table 3 does not state whether the baseline numbers are validation or test mAP, and no confidence intervals or multiple seeds are given. This makes it impossible to judge whether the 5.9-point gap is meaningful or an artifact of validation-based model selection.
  5. [§1, §3.6] The causal attribution to EfficientNet and SE attention is not supported by ablations. No experiment removes the SE blocks, swaps the backbone, or uses the original YOLOv11 backbone/neck on the same data. Given the parameter-count inconsistencies in Table 1, the architecture itself is not uniquely specified, so the statement in the abstract and conclusion that the improvements are due to these components is premature.
minor comments (5)
  1. [Section ordering] The results section jumps from §3.4 to §3.6; §3.5 is missing.
  2. [Table 3] Several entries appear to be transcription errors: Faster R-CNN has identical values for parameters and FLOPs (41.12), and YOLOv11m has identical values for FLOPs and time (29.61). Please check.
  3. [References] In §3.6, reference [40] is cited twice in the same sentence; the second citation should likely be [41] or [42].
  4. [Figure 3] The caption/text says the final mAP of 90.8% was reached after 120 epochs, while Table 2 says training ran for 200 epochs with early stopping disabled. Clarify the relationship between these statements.
  5. [Appendix D] There is a typo: 'Iamge' should be 'Image'.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the reported mAP is an empirical evaluation result, not a quantity forced by the model's definitions or equations.

full rationale

The paper's central claim is an empirical result obtained by training and evaluating a concrete neural network on thermal PV images from Roboflow. The architecture equations (1)-(7) describe standard operations (depthwise/pointwise convolutions, SE recalibration, feature aggregation, YOLO loss, NMS), and Eq. (8) defines mAP as the mean of per-class average precisions; none of these define the reported 90.8% mAP in terms of the model's own design or make the improvement a tautology. Section 2.3 states 'the model achieved a final 𝑚𝐴𝑃 of 90.8% after 120 epochs (using the validation dataset)', and Section 3.6 reports the same value as a measured benchmark result. No fitted parameter is renamed as a prediction, and no baseline number is derived from the proposed model's weights or losses. The paper does cite previous work by the author ([11], [15], [16]), but those references are used as background on PV defect detection and do not carry the burden of justifying the 90.8% mAP, the parameter counts, or the claimed superiority over YOLOv11m. The absence of ablations and the underdocumented baseline training protocol are experimental-rigor concerns that could affect attribution of the improvement to EfficientNet/SE, but they are not circularity: the results are not equal to the inputs by construction. Therefore the paper is self-contained with respect to circularity, and the appropriate score is 0.

Assumptions & free parameters 6 free parameters · 5 assumptions · 0 invented entities

The paper is an empirical ML study rather than a derivation. Its central claim rests on dataset provenance, annotation quality, identical baseline protocols, and the fidelity of the architecture description. The reported hyperparameters are the free parameters of the training setup; no new physical entities are introduced.

free parameters (6)
  • Initial learning rate = 0.001
    Chosen by hand; standard Adam starting point; affects final mAP.
  • Batch size = 16
    Chosen by hand; affects convergence and memory use.
  • Training epochs = 200 (mAP reported at 120)
    Chosen by hand; training is run for 200 epochs but the best validation result is cited at epoch 120.
  • Weight decay = 0.0005
    Regularization constant chosen by hand.
  • Momentum = 0.9
    Optimizer momentum value listed in Table 2.
  • Anchor box sizes = auto-anchor learned from data
    Sizes are fit to the dataset during training and directly affect localization.
assumptions (5)
  • domain assumption The Roboflow-sourced images contain accurate bounding-box annotations and class labels for thermal anomalies.
    Annotations were made with Roboflow tools but no manual verification or inter-annotator agreement is reported (Section 2.3).
  • domain assumption The public Roboflow datasets are representative of drone-based PV thermal inspections.
    The paper does not describe the data provenance, camera, altitude, or weather conditions (Section 2.3).
  • domain assumption All baseline models in Table 3 were trained and evaluated under the same protocol as HOTSPOT-YOLO.
    No baseline training details are provided (Section 3.6).
  • ad hoc to paper The architecture described in Table 1 corresponds to the model that was actually trained.
    Several parameter counts in Table 1 are inconsistent with the listed layer dimensions, so this assumption is not established.
  • domain assumption EfficientNet backbone and SE attention blocks are responsible for the performance gain.
    No ablation isolates either component (Section 3.6).

how reviews work

0 comments
Cite this review

Pith. "Pith review of HOTSPOT-YOLO: A Lightweight Deep Learning Attention-Driven Model for Detecting Thermal Anomalies in Drone-Based Solar Photovoltaic Inspections." pith.science (2026). https://pith.science/paper/P4JQNDGZ

@misc{pith2026250818912,
  author       = {Pith},
  title        = {Pith review of: HOTSPOT-YOLO: A Lightweight Deep Learning Attention-Driven Model for Detecting Thermal Anomalies in Drone-Based Solar Photovoltaic Inspections},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/P4JQNDGZ}},
  note         = {Machine review of arXiv:2508.18912}
}
read the original abstract

Thermal anomaly detection in solar photovoltaic (PV) systems is essential for ensuring operational efficiency and reducing maintenance costs. In this study, we developed and named HOTSPOT-YOLO, a lightweight artificial intelligence (AI) model that integrates an efficient convolutional neural network backbone and attention mechanisms to improve object detection. This model is specifically designed for drone-based thermal inspections of PV systems, addressing the unique challenges of detecting small and subtle thermal anomalies, such as hotspots and defective modules, while maintaining real-time performance. Experimental results demonstrate a mean average precision of 90.8%, reflecting a significant improvement over baseline object detection models. With a reduced computational load and robustness under diverse environmental conditions, HOTSPOT-YOLO offers a scalable and reliable solution for large-scale PV inspections. This work highlights the integration of advanced AI techniques with practical engineering applications, revolutionizing automated fault detection in renewable energy systems.

Figures

Figures reproduced from arXiv: 2508.18912 by the authors.

Figure 1
Figure 1. Conceptual overview of the HOTSPOT-YOLO model. The model builds upon YOLOv11 with significant enhancements, including a lightweight EfficientNet backbone for feature extraction, Squeeze-and-Excitation attention mechanisms to focus on critical regions, improved detection accuracy, and real-time performance optimized for drone-based thermal anomaly detection. 2.2 Detailed Operation of HOTSPOT-YOLO Model The proposed H… view at source ↗
Figure 4
Figure 4. Multi-module hotspot detection results using the HOTSPOT-YOLO model. (a) Detection of a single hotspot in a PV array, with a confidence score of 87%. (b) Detection of a hotspot in a separate PV array, with a confidence score of 90% [PITH_FULL_IMAGE:figures/full_fig_p009_4.png] view at source ↗
Figure 7
Figure 7. Evaluation of HOTSPOT-YOLO under varying image transformations. (a) Original thermal image with two detected hotspots (confidence scores: 74% and 70%). (b) Image with -40% brightness and -40% contrast, where the detection accuracy remains consistent (confidence scores: 73%). (c) Grayscale transformation, showing a minor reduction in confidence scores (68% and 67%), demonstrating the model’s robustness to colour and … view at source ↗
Figures from the paper (3 more)
Figure 8
Figure 8. Figure 8: HOTSPOT-YOLO performance under image blurring conditions. (a) Detection in a highly blurred thermal image with low edge clarity and reduced contrast, identifying hotspots with confidence scores ranging from 53% to 71%. (b) Detection in a moderately blurred image, succe…
Figure 9
Figure 9. Figure 9: HOTSPOT-YOLO detection performance on ground-based thermal images. (a) Wide-angle thermal image with multiple detected hotspots (confidence scores: 63% to 82%), showing limitations in detecting edge anomalies due to angle and focus differences. (b) Close-up thermal ima…
Figure 10
Figure 10. Figure 10: 𝑚𝐴𝑃 progression over epochs for various YOLO models vs. HOTSPOT-YOLO model. The figure highlights the superior performance of HOTSPOT-YOLO with the highest final 𝑚𝐴𝑃@90.8%, demonstrating consistent improvements across training epochs compared to other YOLO variants. F…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

42 extracted references · 40 canonical work pages

  1. [1]

    Cubukcu, M. E. T. E., & Akanalci, A. (2020). Real -time inspection and determination methods of faults on photovoltaic power systems by thermal imaging in Turkey. Renewable Energy, 147, 1231-1238

    work page 2020

  2. [2]

    Zefri, Y ., Sebari, I., Hajji, H., & Aniba, G. (2021). In-depth investigation of applied digital photogrammetry to imagery-based RGB and thermal infrared aerial inspection of large -scale photovoltaic installations. Remote Sensing Applications: Society and Environment, 23, 100576

    work page 2021

  3. [3]

    de Oliveira, A. K. V., Aghaei, M., & Rüther, R. (2022). Automatic inspection of photovoltaic power plants using aerial infrared thermography: a review. Energies, 15(6), 2055

    work page 2022

  4. [4]

    Pruthviraj, U., Kashyap, Y ., Baxevanaki, E., & Kosmopoulos, P . (2023). Solar photovoltaic hotspot inspection using unmanned aerial vehicle thermal images at a solar field in south india. Remote Sensing, 15(7), 1914

    work page 2023

  5. [5]

    Buerhop, C., Bommes , L., Schlipf, J., Pickel, T., Fladung, A., & Peters, I. M. (2022). Infrared imaging of photovoltaic modules: a review of the state of the art and future challenges facing gigawatt photovoltaic power stations. Progress in Energy, 4(4), 042010

    work page 2022

  6. [6]

    Dhimish, M., & Badran, G. (2023). Investigating defects and annual degradation in UK solar PV installations through thermographic and electroluminescent surveys. npj Materials Degradation, 7(1), 14

    work page 2023

  7. [7]

    A., Urmee, T., & Parlevliet, D

    Rahaman, S. A., Urmee, T., & Parlevliet, D. A. (2022). Investigate the impact of environmental and operating conditions of infrared (IR) imaging on infrared thermography of PV modules to identify defects. Solar Energy, 245, 231-253

    work page 2022

  8. [8]

    G., Burud, I., Olsen, E., & Belbachir, N

    Høiaas, I., Grujic, K., Imenes, A. G., Burud, I., Olsen, E., & Belbachir, N. (2022). Inspection and condition monitoring of large -scale photovoltaic power plants: A review of imaging technologies. Renewable and Sustainable Energy Reviews, 161, 112353

    work page 2022

Show all 42 references
  1. [9]

    Yahya, Z., Imane, S., Hicham, H., Ghassane, A., & Safia, E. B. I. (2022). Applied imagery pattern recognition for photovoltaic modules’ inspection: A review on methods, challenges and future development. Sustainable Energy Technologies and Assessments, 52, 102071

    2022

  2. [10]

    B., Leon -Salas, W

    Bosman, L. B., Leon -Salas, W. D., Hutzel, W., & Soto, E. A. (2020). PV system predictive maintenance: Challenges, current approaches, and opportunities. Energies, 13(6), 1398

    2020

  3. [11]

    and Dhimish, M., 2023

    Hassan, S. and Dhimish, M., 2023. Dual spin max pooling convolutional neural network for solar cell crack detection. Scientific reports, 13(1), p.11099

    2023

  4. [12]

    and Dhimish, M., 2025

    Drir, N., Mellit, A., Bettayeb, M. and Dhimish, M., 2025. Enhanced Photovoltaic Defect Detection Using Perceptual Loss in DCGAN and VGG16 -Integrated Models on Electroluminescence Images. IEEE Journal of Photovoltaics.sssss

    2025

  5. [13]

    H., Marugán, A

    Herraiz, Á. H., Marugán, A. P ., & Márquez, F. P . G. (2020). Photovoltaic plant condition monitoring using thermal images analysis by convolutional neural network -based structure. Renewable Energy, 153, 334-348

    2020

  6. [14]

    F., Badjugar, C., Nepal, U., Poulose, A.,

    Sapkota, R., Qureshi, R., Calero, M. F., Badjugar, C., Nepal, U., Poulose, A., ... & Karkee, M. (2024). YOLOv10 to Its Genesis: A Decadal and Comprehensive Review of The You Only Look Once (YOLO) Series. arXiv preprint arXiv:2406.19407

    2024 arXiv

  7. [15]

    (2024, November)

    Aktouf, L., Shivanna, Y ., & Dhimish, M. (2024, November). High-Precision Defect Detection in Solar Cells Using YOLOv10 Deep Learning Model. In Solar (Vol. 4, No. 4, pp. 639-659). MDPI

    2024

  8. [16]

    (2023, December)

    Hassan, S., & Dhimish, M. (2023, December). A Survey of CNN -Based Approaches for Crack Detection in Solar PV Modules: Current Trends and Future Directions. In Solar (Vol. 3, No. 4, pp. 663-683). MDPI

    2023

  9. [17]

    (2024, June)

    Hussain, M., & Khanam, R. (2024, June). In -depth review of yolov1 to yolov10 variants for enhanced photovoltaic defect detection. In Solar (Vol. 4, No. 3, pp. 351-386). MDPI. Page 28 of 29

    2024

  10. [18]

    Khanam, R., & Hussain, M. (2024). YOLOv11: An Overview of the Key Architectural Enhancements. arXiv preprint arXiv:2410.17725

    2024 arXiv

  11. [19]

    Yang, X., Li, Y ., Yang, L., Zhang, Y ., Wang, X., & Zhang, Q. (2024). High-noise solar panel defect identification method based on the improved EfficientNet -V2. Journal of Renewable and Sustainable Energy, 16(5)

    2024

  12. [20]

    Liu, B., Chen, L., Sun, K., Wang, X., & Zhao, J. (2024). A Hot Spot Identification Approach for Photovoltaic Module Based on Enhanced U-Net With Squeeze-and-Excitation and VGG19. IEEE Transactions on Instrumentation and Measurement

    2024

  13. [21]

    S., Zhao, B

    Jin, X., Xie, Y ., Wei, X. S., Zhao, B. R., Chen, Z. M., & Tan, X. (2022). Delving deep into spatial pooling for squeeze-and-excitation networks. Pattern Recognition, 121, 108159

    2022

  14. [22]

    Qi, Q., Zhao, J., Lin, L., Zhang, X., & Tian, Y . (2024). Combined multi-level context aggregation and attention mechanism method for photovoltaic panel extraction from high resolution remote sensing images. International Journal of Remote Sensing, 45(11), 3560-3576

    2024

  15. [23]

    Gong, B., An, A., Shi, Y ., & Zhang, X. (2024). Fast fault detection method for photovoltaic arrays with adaptive deep multiscale feature enhancement. Applied Energy, 353, 122071

    2024

  16. [24]

    Wang, Y ., Hou, T., Zhang, X., Shangguan, H., Zhang, P ., Li, J., & Wei, B. (2023). Surface defect detection of solar cell based on similarity non -maximum suppression mechanism. Signal, Image and Video Processing, 17(5), 2583-2593

    2023

  17. [25]

    Mazen, F. M. A., Seoud, R. A. A., & Shaker, Y . O. (2023). Deep learning for automatic defect detection in PV modules using electroluminescence images. IEEE Access, 11, 57783-57795

    2023

  18. [26]

    Mahasin, M., & Dewi, I. A. (2022). Comparison of CSPDarkNet53, CSPResNeXt -50, and EfficientNet-B0 backbones on YOLO v4 as object detector. International journal of engineering, science and information technology, 2(3), 64-72

    2022

  19. [27]

    Jiang, P ., Ergu, D., Liu, F., Cai, Y ., & Ma, B. (2022). A Review of Yolo algorithm developments. Procedia computer science, 199, 1066-1073

    2022

  20. [28]

    Wang, Z., Zhou, D., Guo, C., & Zhou, R. (2024). Yolo -global: a real -time target detector for mineral particles. Journal of Real-Time Image Processing, 21(3), 1-13

    2024

  21. [29]

    R., Singh, S

    Dubey, S. R., Singh, S. K., & Chaudhuri, B. B. (2022). Activation functions in deep learning: A comprehensive survey and benchmark. Neurocomputing, 503, 92-108

    2022

  22. [30]

    Apicella, A., Donnarumma, F., Isgrò, F., & Prevete, R. (2021). A survey on modern trainable activation functions. Neural Networks, 138, 14-32

    2021

  23. [31]

    Ni, X., Ma, Z., Liu, J., Shi, B., & Liu, H. (2021). Attention network for rail surface defect detection via consistency of intersection -over-union (IoU) -guided center -point estimation. IEEE Transactions on Industrial Informatics, 18(3), 1694-1705

    2021

  24. [32]

    Soydaner, D. (2020). A comparison of optimization algorithms for deep learning. International Journal of Pattern Recognition and Artificial Intelligence, 34(13), 2052013

    2020

  25. [33]

    M., Schneider, F., & Hennig, P

    Schmidt, R. M., Schneider, F., & Hennig, P . (2021, July). Descending through a crowded valley- benchmarking deep learning optimizers. In International Conference on Machine Learning (pp. 9367-9376). PMLR

    2021

  26. [34]

    Jia, Y ., Chen, G., & Zhao, L. (2024). Defect detection of photovoltaic modules based on improved VarifocalNet. Scientific Reports, 14(1), 15170

    2024

  27. [35]

    Xia, H., Yang, B., Li, Y ., & Wang, B. (2022). An improved CenterNet model for insulator defect detection using aerial imagery. Sensors, 22(8), 2850

    2022

  28. [36]

    (2023, May)

    Wang, J., Zhang, R., & Zheng, X. (2023, May). Photovoltaic Panel Intelligent Detection Method Based on Improved Faster -RCNN. In 2023 IEEE 3rd International Conference on Electronic Technology, Communication and Information (ICETCI) (pp. 1565-1569). IEEE

    2023

  29. [37]

    Link: https://docs.ultralytics.com/models/yolov5/#overview Page 29 of 29

    Ultralytics YOLOv5 (2023). Link: https://docs.ultralytics.com/models/yolov5/#overview Page 29 of 29

    2023

  30. [38]

    Link: https://docs.ultralytics.com/models/yolov9/#what-tasks-and- modes-does-yolov9-support

    Ultralytics YOLOv9 (2024). Link: https://docs.ultralytics.com/models/yolov9/#what-tasks-and- modes-does-yolov9-support

    2024

  31. [39]

    Link: https://docs.ultralytics.com/models/yolo11/

    Ultralytics YOLOv11 (2024). Link: https://docs.ultralytics.com/models/yolo11/

    2024

  32. [40]

    Han, X., Wang, X., Chen, C., Li, G., & Piao, C. (2023). Hot Spot Detection of Thermal Infrared Image of Photovoltaic Power Station Based on Multi -Task Fusion. Journal of Information Processing Systems, 19(6), 791-802

    2023

  33. [41]

    A., & Dawood, Z

    Numan, A., Hussein, H. A., & Dawood, Z. S. (2021). Hot spot analysis of photovoltaic module under partial shading conditions by using IR-imaging technology. Engineering and Technology Journal, 39(9), 1338-1344

    2021

  34. [42]

    U., Saleem, S., Masood, H., Kallu, K

    Ali, M. U., Saleem, S., Masood, H., Kallu, K. D., Masud, M., Alvi, M. J., & Zafar, A. (2022). Early hotspot detection in photovoltaic modules using color image descriptors: An infrared thermography study. International Journal of Energy Research, 46(2), 774-785

    2022

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.