Computer Vision-Based Automated Inspection of Solar Panels using Drone Imagery and  Deep Learning

Paul Praneeth Muddamalle; Paul Praneeth Muddamalle

doi:10.15662/ijetemr.2025.1001001

Computer Vision-Based Automated Inspection of Solar Panels using Drone Imagery and Deep Learning

Paul Praneeth Muddamalle

Golden Gate University, USA

Abstract

Utility-scale photovoltaic farms need inspection methods that can do more than flag a defective panel. The system must identify the defect, locate the affected area, and give maintenance teams enough information to decide which panels require attention first. Drone-based inspection has made large-scale photovoltaic monitoring more feasible, yet many automated models still rely on a single image modality, produce only classification outputs, receive limited robustness testing, or remain weakly connected to maintenance planning. To address these gaps, an explainable RGB-thermal drone-vision framework was developed and validated for localization-aware and severity-ranked solar panel defect inspection in utility-scale photovoltaic farms. The study used 21,500 paired RGB-thermal UAV images collected from operational solar farms over six months. The dataset included seven defect categories and normal panels. Three solar inspection experts annotated the images with bounding boxes, segmentation masks, and severity labels, which allowed the model to be evaluated not only for recognition accuracy but also for spatial localization and maintenance relevance. The proposed dual-branch architecture combines RGB feature extraction, thermal feature extraction, multimodal fusion attention, defect detection, segmentation, severity ranking, and explainability assessment. On the internal test set, the model achieved 96.3% precision, 94.8% recall, 95.5% F1-score, 96.9% mAP@0.5, 83.7% mAP@0.5:0.95, and 84.9% IoU. External validation on an unseen solar farm produced a 92.8% F1-score, which suggests that the model retained good performance beyond the original data collection sites. The explainability module reached 86.7% expert-mask overlap, while severity classification achieved a 92.5% F1-score and maintenance-priority classification reached 94.2% accuracy. Overall, the results indicate that RGB-thermal fusion provides stronger defect detection, localization, interpretability, and maintenance-oriented decision support than RGB-only, thermal-only, and conventional baseline models

Article Information

Journal	International Journal of Emerging Trends in Engineering and Management Research
Volume (Issue)	Vol. 10 No. 1 (2025): International Journal of Emerging Trends in Engineering and Management Research (IJETEMR)
DOI	https://doi.org/10.15662/ijetemr.2025.1001001
Pages	17091-17106
Published	January 13, 2025
Copyright	All rights reserved
Open Access	This work is licensed under a Creative Commons Attribution 4.0 International License.
How to Cite	Paul Praneeth Muddamalle (%2025). Computer Vision-Based Automated Inspection of Solar Panels using Drone Imagery and Deep Learning. International Journal of Emerging Trends in Engineering and Management Research , Vol. 10 No. 1 (2025): International Journal of Emerging Trends in Engineering and Management Research (IJETEMR) , pp. 17091-17106. https://doi.org/10.15662/ijetemr.2025.1001001

References

[1] International Energy Agency, Renewables 2023: Analysis and Forecast to 2028. Paris: IEA, 2023.
[2] U. Jahn, M. Herz, M. Köntges, D. Parlevliet, T. Paggi, I. Tsanakas, and S. Stein, Review of Failures of Photovoltaic Modules. IEA-PVPS Task 13, Report IEA-PVPS T13-01:2014, 2014.
[3] D. C. Jordan and S. R. Kurtz, “Photovoltaic degradation rates: An analytical review,” Progress in Photovoltaics: Research and Applications, vol. 21, no. 1, pp. 12–29, 2013. DOI: 10.1002/pip.1182
[4] A. Ndiaye, A. Charki, A. Kobi, C. M. F. Kébé, P. A. Ndiaye, and V. Sambou, “Degradations of silicon photovoltaic modules: A literature review,” Solar Energy, vol. 96, pp. 140–151, 2013. DOI: 10.1016/j.solener.2013.07.005
[5] I. Tsanakas, J. Ha, and C. Buerhop, “Faults and infrared thermographic diagnosis in operating c-Si photovoltaic modules: A review of research and future challenges,” Renewable and Sustainable Energy Reviews, vol. 62, pp. 695–709, 2016. DOI: 10.1016/j.rser.2016.04.079
[6] C. Buerhop, D. Schlegel, M. Niess, C. Vodermayer, R. Weißmann, and C. J. Brabec, “Reliability of IR-imaging of PV-plants under operating conditions,” Solar Energy Materials and Solar Cells, vol. 107, pp. 154–164, 2012. DOI: 10.1016/j.solmat.2012.07.011
[7] S. A. A. Madeti and S. N. Singh, “A comprehensive study on different types of faults and detection techniques for solar photovoltaic system,” Solar Energy, vol. 158, pp. 161–185, 2017. DOI: 10.1016/j.solener.2017.08.069
[8] F. Grimaccia, M. Aghaei, M. Mussetta, S. Leva, and P. B. Quater, “Planning for PV plant performance monitoring by means of unmanned aerial systems,” International Journal of Energy and Environmental Engineering, vol. 6, pp. 47–54, 2015. DOI: 10.1007/s40095-014-0149-6
[9] M. Aghaei, F. Grimaccia, C. A. Gonano, and S. Leva, “Innovative automated control system for PV fields inspection and remote control,” IEEE Transactions on Industrial Electronics, vol. 62, no. 11, pp. 7287–7296, 2015. DOI: 10.1109/TIE.2015.2437452
[10] M. Köntges, S. Kurtz, C. Packard, U. Jahn, K. Berger, K. Kato, T. Friesen, H. Liu, and M. Van Iseghem, Performance and Reliability of Photovoltaic Systems: Review of Failures of Photovoltaic Modules. IEA-PVPS Task 13, 2014.
[11] G. T. Puliti, G. Calamita, F. Cigna, and M. Tapete, “Unmanned aerial vehicle thermal infrared imaging for monitoring photovoltaic plants: A review,” Remote Sensing, vol. 13, no. 10, 2021.
[12] A. Mellit, G. M. Tina, and S. A. Kalogirou, “Fault detection and diagnosis methods for photovoltaic systems: A review,” Renewable and Sustainable Energy Reviews, vol. 91, pp. 1–17, 2018. DOI: 10.1016/j.rser.2018.03.062
[13] H. Mekki, A. Mellit, and H. Salhi, “Artificial neural network-based modelling and fault detection of partial shaded photovoltaic modules,” Simulation Modelling Practice and Theory, vol. 67, pp. 1–13, 2016.
[14] A. K. S. Jardine, D. Lin, and D. Banjevic, “A review on machinery diagnostics and prognostics implementing condition-based maintenance,” Mechanical Systems and Signal Processing, vol. 20, no. 7, pp. 1483–1510, 2006. DOI: 10.1016/j.ymssp.2005.09.012
[15] R. K. Mobley, An Introduction to Predictive Maintenance, 2nd ed. Boston: Butterworth-Heinemann, 2002.
[16] S. Deitsch, C. Buerhop-Lutz, E. Sovetkin, A. Steland, F. A. Maier, and C. Riess, “Segmentation of photovoltaic module cells in uncalibrated electroluminescence images,” Machine Vision and Applications, vol. 32, no. 84, 2021.
[17] S. Deitsch, C. Buerhop-Lutz, E. Sovetkin, A. Steland, F. A. Maier, A. Gallwitz, and C. Riess, “Automatic classification of defective photovoltaic module cells in electroluminescence images,” Solar Energy, vol. 185, pp. 455–468, 2019. DOI: 10.1016/j.solener.2019.02.067
[18] R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh, and D. Batra, “Grad-CAM: Visual explanations from deep networks via gradient-based localization,” in Proceedings of the IEEE International Conference on Computer Vision, 2017, pp. 618–626. DOI: 10.1109/ICCV.2017.74
[19] A. Adadi and M. Berrada, “Peeking inside the black-box: A survey on explainable artificial intelligence,” IEEE Access, vol. 6, pp. 52138–52160, 2018. DOI: 10.1109/ACCESS.2018.2870052
[20] J. Lee, F. Wu, W. Zhao, M. Ghaffari, L. Liao, and D. Siegel, “Prognostics and health management design for rotary machinery systems: Reviews, methodology and applications,” Mechanical Systems and Signal Processing, vol. 42, no. 1–2, pp. 314–334, 2014. DOI: 10.1016/j.ymssp.2013.06.004
[21] Y. LeCun, Y. Bengio, and G. Hinton, “Deep learning,” Nature, vol. 521, pp. 436–444, 2015. DOI: 10.1038/nature14539
[22] K. He, X. Zhang, S. Ren, and J. Sun, “Deep residual learning for image recognition,” in Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2016, pp. 770–778. DOI: 10.1109/CVPR.2016.90
[23] J. Redmon and A. Farhadi, “YOLOv3: An incremental improvement,” arXiv preprint arXiv:1804.02767, 2018.
[24] A. Bochkovskiy, C. Y. Wang, and H. Y. M. Liao, “YOLOv4: Optimal speed and accuracy of object detection,” arXiv preprint arXiv:2004.10934, 2020.
[25] Z. Ge, S. Liu, F. Wang, Z. Li, and J. Sun, “YOLOX: Exceeding YOLO series in 2021,” arXiv preprint arXiv:2107.08430, 2021.
[26] O. Ronneberger, P. Fischer, and T. Brox, “U-Net: Convolutional networks for biomedical image segmentation,” in Medical Image Computing and Computer-Assisted Intervention, 2015, pp. 234–241. DOI: 10.1007/978-3-319-24574-4_28
[27] E. Trist, “The evolution of socio-technical systems,” Occasional Paper No. 2, Ontario Quality of Working Life Centre, 1981.
[28] A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, Ł. Kaiser, and I. Polosukhin, “Attention is all you need,” in Advances in Neural Information Processing Systems, 2017, pp. 5998–6008.
[29] T. Lin, P. Goyal, R. Girshick, K. He, and P. Dollár, “Focal loss for dense object detection,” in Proceedings of the IEEE International Conference on Computer Vision, 2017, pp. 2980–2988. DOI: 10.1109/ICCV.2017.324
[30] M. Everingham, L. Van Gool, C. K. I. Williams, J. Winn, and A. Zisserman, “The Pascal Visual Object Classes challenge: A retrospective,” International Journal of Computer Vision, vol. 111, pp. 98–136, 2015. DOI: 10.1007/s11263-014-0733-5