Real-time detection method for photovoltaic electroluminescence images targeting weak and multi-scale defects

i

This article has been updated. A Correction to this article was published on .

i

The Original Article was published on

i

The Original Article was published on

i

This article has been updated

i

This article was retracted on

i

This article has been updated

Abstract

Defect detection in photovoltaic (PV) modules is crucial for long-term reliable operation of solar plants, reducing levelized cost of energy, and preventing safety incidents. Electroluminescence (EL) images suffer from low contrast, complex background noise, and frequent overlapping defects, hindering the balance between detection accuracy and real-time processing speed for current algorithms. To address this, a novel real-time object detector is proposed. First, a convolutional additive self-attention block (CAS-Block) reconstructs the backbone by replacing dot-product with additive similarity, enhancing the response sensitivity to weak defect features in low-contrast images while maintaining linear computational complexity. Second, an efficient dynami-scale intra-feature interaction (EDII) module based on frequency-domain learning employs adaptive spectral filtering to suppress environmental noise interference, such as illumination variations. Finally, a gated cross-scale feature fusion (GCFF) module built on gated linear unit (GLU) uses a dynamic gating mechanism to precisely disentangle feature confusion arising from overlapping defects. Experimental results show that the proposed model achieves a mean average precision (mAP)@0.5 of 93.2% and a detection speed of 104.2 frames per second (FPS), outperforming the state-of-the-art (SOTA) YOLOv12-L and EER- DETR. The method achieves an optimal accuracy–speed trade-off, offering an efficient, cost-effective solution for intelligent inspection of large-scale PV plants.

Keywords

photovoltaic cells / defect detection / electroluminescence images / deep learning / real-time object detection

CHANGE HISTORY

A Correction to this paper has been published:
https://doi.org

The Retraction Note was updated to reflect reception of statements of disagreement from all authors.

This article has been retracted.Please see the Retraction Notice for more detail:
https://doi.org

References

[1] B. Su, Z. Zhou, H. Chen, PVEL-AD: A large-scale open-world dataset for photovoltaic cell anomaly detection, IEEE Trans. Ind. Inform., 19 (2023), 404-413. https://doi.org/10.1109/TII.2022.3162846

[2] V. Sharma, S. Chandel, Performance and degradation analysis for long term reliability of solar photovoltaic systems: A review, Renew. Sust. Energ. Rev., 27 (2013), 753-767. https://doi.org/10.1016/j.rser.2013.07.046

[3] S. Gallardo-Saavedra, L. Hernández-Callejo, M. Alonso-García, J. Santos, J. Morales-Aragonés, V. Alonso-Gómez, et al., Nondestructive characterization of solar PV cells defects by means of electroluminescence, infrared thermography, I-V curves and visual tests: Experimental study and comparison, Energy, 205 (2020), 117930. https://doi.org/10.1016/j.energy.2020.117930

[4] S. A. Memon, Q. Javed, W. G. Kim, Z. Mahmood, U. Khan, M. Shahzad, A machine-learning-based robust classification method for PV panel faults, Sensors, 22 (2022), 8515. https://doi.org/10.3390/s22218515

[5] J. Fioresi, D. J. Colvin, R. Frota, R. Gupta, M. Li, H. P. Seigneur, et al., Automated defect detection and localization in photovoltaic cells using semantic segmentation of electroluminescence images, IEEE J. Photovolt., 12 (2022), 53-61. https://doi.org/10.1109/JPHOTOV.2021.3131059

[6] J. Tian, C. Chen, W. Shen, F. Sun, R. Xiong, Deep learning framework for lithium-ion battery state of charge estimation: Recent advances and future perspectives, Energy Storage Mater., 61 (2023), 102883. https://doi.org/10.1016/j.ensm.2023.102883

[7] H. Tella, A. Hussein, S. Rehman, B. Liu, A. Balghonaim, M. Mohandes, Solar photovoltaic panel cells defects classification using deep learning ensemble methods, Case Stud. Therm. Eng., 66 (2025), 105749. https://doi.org/10.1016/j.csite.2025.105749

[8] S. Shah, D. Suraj, S. M. Reza, M. A. R. B. A. Salam, A. Ashraf, S. F. Ferdous, Comparative analysis of deep learning models for defect detection in additive manufacturing using thermal imaging, Results Eng., 28 (2025), 108359. https://doi.org/10.1016/j.rineng.2025.108359

[9] A. Krizhevsky, I. Sutskever, G. E. Hinton, ImageNet classification with deep convolutional neural networks, in Advances in Neural Information Processing Systems 25, Curran Associates, Inc., (2012), 1097-1105.

[10] K. Simonyan, A. Zisserman, Very deep convolutional networks for large-scale image recognition, preprint, arXiv:1409.1556.

[11] K. He, X. Zhang, S. Ren, J. Sun, Deep residual learning for image recognition, in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, (2016), 770-778. https://doi.org/10.1109/CVPR.2016.90

[12] M. Sandler, A. G. Howard, M. Zhu, A. Zhmoginov, L. Chen, MobileNetV2: Inverted residuals and linear bottlenecks, in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, (2018), 4510-4520. https://doi.org/10.1109/CVPR.2018.00474

[13] W. Tang, Q. Yang, X. Hu, W. Yan, Deep learning-based linear defects detection system for large-scale photovoltaic plants based on an edge-cloud computing infrastructure, Sol. Energy, 231 (2022), 527-535. https://doi.org/10.1016/j.solener.2021.11.016

[14] A. Howard, M. Sandler, B. Chen, W. Wang, L. C. Chen, M. Tan, et al., Searching for MobileNetV3, in IEEE/CVF International Conference on Computer Vision (ICCV), IEEE, (2019), 1314-1324. https://doi.org/10.1109/ICCV.2019.00140

[15] L. Pratt, D. Govender, R. Klein, Defect detection and quantification in electroluminescence images of solar PV modules using U-net semantic segmentation, Renew. Energy, 178 (2021), 1211-1222. https://doi.org/10.1016/j.renene.2021.06.086

[16] U. Hijjawi, S. Lakshminarayana, T. Xu, G. P. M. Fierro, M. Rahman, A review of automated solar photovoltaic defect detection systems: Approaches, challenges, and future orientations, Sol. Energy, 266 (2023), 112186. https://doi.org/10.1016/j.solener.2023.112186

[17] A. Vaswani, N. Shazeer, N. Parmar, J. Uszkoreit, L. Jones, A. N. Gomez, et al., Attention is all you need, in Advances in Neural Information Processing Systems 30 (NeurIPS 2017), Curran Associates, Inc., (2017), 5998-6008.

[18] N. Carion, F. Massa, G. Synnaeve, N. Usunier, A. Kirillov, S. Zagoruyko, End-to-end object detection with transformers, in Proceedings of the European Conference on Computer Vision (ECCV), Springer, (2020), 213-229. https://doi.org/10.1007/978-3-030-58452-813

[19] Z. Liu, Y. Lin, Y. Cao, H. Hu, Y. Wei, Z. Zhang, et al., Swin transformer: Hierarchical vision transformer using shifted windows, in IEEE/CVF International Conference on Computer Vision (ICCV), IEEE, (2021), 9992-10002. https://doi.org/10.1109/ICCV48922.2021.00986

[20] H. Zhang, F. Li, S. Liu, L. Zhang, H. Su, J. Zhu, et al., DINO: DETR with improved denoising anchor boxes for end-to-end object detection, preprint, arXiv:2203.03605.

[21] Y. Zhao, W. Lv, S. Xu, J. Wei, G. Wang, Q. Dang, et al., DETRs beat YOLOs on real-time object detection, in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, (2024), 16965-16974. https://doi.org/10.1109/CVPR52733.2024.01605

[22] J. Dun, H. Yang, S. Yuan, Y. Tang, EER-DETR: An improved method for detecting defects on the surface of solar panels based on RT-DETR, Appl. Sci., 15 (2025), 6217. https://doi.org/10.3390/app15116217

[23] M. W. Akram, G. Li, Y. Jin, CNN based automatic detection of photovoltaic cell defects in electroluminescence images, Energy, 189 (2019), 116319. https://doi.org/10.1016/j.energy.2019.116319

[24] M. Y. Demirci, N. Besli, A. Gümüşcü, Efficient deep feature extraction and classification for identifying defective photovoltaic module cells in electroluminescence images, Expert Syst. Appl., 175 (2021), 114810. https://doi.org/10.1016/j.eswa.2021.114810

[25] J. Zhang, Y. Wang, B. Jiang, H. He, S. Huang, C. Wang, et al., Realistic fault detection of Li-ion battery via dynamical deep learning, Nat. Commun., 14 (2023), 5940. https://doi.org/10.1038/s41467-023-41226-5

[26] Q. Liu, M. Liu, C. Wang, Q. M. Wu, An efficient CNN-based detector for photovoltaic module cells defect detection in electroluminescence images, Sol. Energy, 267 (2024), 112245. https://doi.org/10.1016/j.solener.2023.112245

[27] T. Zhang, L. Li, Y. Zhou, W. Liu, C. Qian, J. N. Hwang, et al., CAS-ViT: Convolutional additive self-attention vision transformers for efficient mobile applications, IEEE Trans. Image Process., 35 (2026), 1899-1909. https://doi.org/10.1109/TIP.2026.3655121

[28] Y. Rao, W. Zhao, Z. Zhu, J. Lu, J. Zhou, Global filter networks for image classification, in Advances in Neural Information Processing Systems 34 (NeurIPS 2021), Curran Associates, Inc., (2021), 980-993.

[29] L. Kong, J. Dong, J. Ge, M. Li, J. Pan, Efficient frequency domain-based transformers for high-quality image deblurring, in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, (2023), 5886-5895. https://doi.org/10.1109/CVPR52729.2023.00570

[30] Y. N. Dauphin, A. Fan, M. Auli, D. Grangier, Language modeling with gated convolutional networks, in Proceedings of the 34th International Conference on Machine Learning, PMLR, (2017), 933-941.

[31] C. Y. Wang, H. Y. M. Liao, Y. H. Wu, P. Y. Chen, J. W. Hsieh, I. H. Yeh, CSPNet: A new backbone that can enhance learning capability of CNN, in IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW), IEEE, (2020), 390-391. https://doi.org/10.1109/CVPRW50498.2020.00203

[32] A. Mohamed, Y. Nacera, B. Ahcene, A. Teta, E. O. Belabbaci, A. Rabehi, et al., Optimized YOLO based model for photovoltaic defect detection in electroluminescence images, Sci. Rep., 15 (2025), 32955. https://doi.org/10.1038/s41598-025-13956-7

[33] G. Jocher, A. Chaurasia, A. Stoken, J. Borovec, NanoCode012, Y. Kwon, et al., Ultralytics YOLOv5 (v7.0)-YOLOv5 sota realtime instance segmentation, 2022. https://doi.org/10.5281/zenodo.7347926

[34] C. Li, L. Li, H. Jiang, K. Weng, Y. Geng, L. Li, et al., YOLOv6: A single-stage object detection framework for industrial applications, preprint, arXiv:2209.02976.

[35] C. Y. Wang, A. Bochkovskiy, H. Y. M. Liao, YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors, in IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), IEEE, (2023), 7464-7475. https://doi.org/10.1109/CVPR52729.2023.00721

[36] G. Jocher, A. Chaurasia, J. Qiu, Ultralytics YOLOv8 (v8.0), 2023. Available from: https://github.com/ultralytics/ultralytics.

[37] C. Y. Wang, I. H. Yeh, H. Y. M. Liao, YOLOv9: Learning what you want to learn using programmable gradient information, in Lecture Notes in Computer Science, Springer, (2024), 1-21. https://doi.org/10.1007/978-3-031-72751-11

[38] A. Wang, H. Chen, L. Liu, K. Chen, Z. Lin, J. Han, et al., YOLOv10: Real-time end-to-end object detection, in Advances in Neural Information Processing Systems 37 (NeurIPS 2024), Curran Associates, Inc., (2024), 107984-108011.

[39] R. Khanam, M. Hussain, YOLOv11: An overview of the key architectural enhancements, preprint, arXiv:2410.17725.

[40] Y. Tian, Q. Ye, D. Doermann, YOLOv12: Attention-centric real-time object detectors, in Advances in Neural Information Processing Systems 38 (NeurIPS 2025), Curran Associates, Inc., (2025), 78433-78457.