A Lightweight Convnext-Based Architecture with Hybrid Feature Aggregation for Multiclass Photovoltaic Panel Defect Classification

Abstract

Photovoltaic panel defect classification is essential for reliable solar energy systems, particularly in industrial inspection. This study proposes a lightweight ConvNeXt-based architecture with a hybrid feature aggregation mechanism that combines Global Average Pooling and Global Max Pooling, followed by a compact classification head for multiclass defect classification. Unlike standard pooling strategies, the proposed aggregation enhances feature representation by capturing both global distribution and salient responses. The proposed model is evaluated through ablation studies and compared with baseline configurations under a unified training pipeline. Experimental results show that the proposed architecture achieves a weighted F1-score of approximately 0.96, outperforming baseline ConvNeXt variants and demonstrating improved feature representation capability. These findings indicate that effective yet simple feature aggregation strategies can significantly enhance classification performance while retaining reasonable inference speed for industrial applications, making the model suitable for real-world PV inspection applications.

Keywords

• Photovoltaic panel defect classification
• ConvNeXt
• Transfer learning
• Feature aggregation
• Deep learning

References

1. Ahmed, W., Hanif, A., Kallu, K. D., Kouzani, A. Z., Ali, M. U., & Zafar, A. (2021). Photovoltaic panels classification using isolated and transfer learned deep neural models using infrared thermographic images. Sensors, 21(16), Article 5668. https://doi.org/10.3390/s21165668
2. Akram, M. W., & Bai, J. (2025). Defect detection in photovoltaic modules based on image-to-image generation and deep learning. Sustainable Energy Technologies and Assessments, 82, Article 104441. https://doi.org/10.1016/j.seta.2025.104441
3. Cao, Y., Pang, D., Yan, Y., Jiang, Y., & Tian, C. (2023). A photovoltaic surface defect detection method for building based on deep learning. Journal of Building Engineering, 70, Article 106375. https://doi.org/10.1016/j.jobe.2023.106375
4. Demir, F. (2025). Enhancing defect classification in solar panels with electroluminescence imaging and advanced machine learning strategies. IEEE Access, 13, 58481–58495. https://doi.org/10.1109/ACCESS.2025.3551749
5. Diaconu, B. M. (2026). Diagnosing shortcut learning in CNN-based photovoltaic fault recognition from RGB images: A multi-method explainability audit. AI, 7(3), Article 94. https://doi.org/10.3390/ai7030094
6. Ejiyi, C. J., Cai, D., Johnson, N., Osei-Mensah, E., Eze, F., Asare, S. K., Staffell, I., & Bamisile, O. O. (2026). SolarSynthNet (SSN): A deep learning framework for binary and multiclass classification of damaged or obstructed solar panels using images. Renewable Energy, 256, Article 124224. https://doi.org/10.1016/j.renene.2025.124224
7. Eren, B. (2026). Deep learning approaches for weld defect detection: A comprehensive review of models, applications, and future directions. Computers & Industrial Engineering, 212, Article 111725. https://doi.org/10.1016/j.cie.2025.111725
8. Hijjawi, U., Lakshminarayana, S., Xu, T., Piero Malfense Fierro, G., & Rahman, M. (2023). A review of automated solar photovoltaic defect detection systems: Approaches, challenges, and future orientations. Solar Energy, 266, Article 112186. https://doi.org/10.1016/j.solener.2023.112186

9. Huang, J., Ariffin, S. A., Chen, Y., Lin, J., & Xu, W. (2025). A novel MoCo-based self-supervised learning framework for solar panel defect detection. IEEE Access, 13, 22977–22988. https://doi.org/10.1109/ACCESS.2025.3529701
10. Ibn Malek, I., & Imtiaz, H. (2025). Advanced fault detection in PV panels using deep neural networks: Leveraging transfer learning and electroluminescence image processing. Energy Advances, 180–193. https://doi.org/10.1039/d5ya00239g
11. Jiang, Y., & Zhao, C. (2022). Attention classification-and-segmentation network for micro-crack anomaly detection of photovoltaic module cells. Solar Energy, 238, 291–304. https://doi.org/10.1016/j.solener.2022.04.012
12. Khanam, R., Hussain, M., Hill, R., & Allen, P. (2024). A comprehensive review of convolutional neural networks for defect detection in industrial applications. IEEE Access, 12, 94250–94295. https://doi.org/10.1109/ACCESS.2024.3425166
13. Laouamer, M., Adaika, M., Remha, S., Mahmoudi, A., & Adaika, H. (2025). Comparative deep learning for RGB-based PV surface fault classification using ResNet50 and EfficientNetB0 with real-time deployment. International Journal of Electrical and Electronics Research, 13(4), Article 712. https://doi.org/10.37391/ijeer.130411
14. Ledmaoui, Y., El Maghraoui, Adila., El Aroussi, M., & Saadane, R. (2024). Enhanced fault detection in photovoltaic panels using CNN-based classification with PyQt5 implementation. Sensors, 24(22), Article 7407. https://doi.org/10.3390/s24227407
15. Masita, K., Hasan, A., Shongwe, T., & Hilal, H. A. (2025). Deep learning in defects detection of PV modules: A review. Solar Energy Advances, 5, Article 100090. https://doi.org/10.1016/j.seja.2025.100090
16. Munawer Al-Otum, H. (2024). Classification of anomalies in electroluminescence images of solar PV modules using CNN-based deep learning. Solar Energy, 278, Article 112803. https://doi.org/10.1016/j.solener.2024.112803
17. Pathak, S. P., & Patil, S. A. (2023). Evaluation of effect of pre-processing techniques in solar panel fault detection. IEEE Access, 11, 72848–72860. https://doi.org/10.1109/ACCESS.2023.3293756
18. Shaban, W. M. (2024). Detection and classification of photovoltaic module defects based on artificial intelligence. Neural Computing and Applications, 36(27), 16769–16796. https://doi.org/10.1007/s00521-024-10000-z
19. Sun, Y., Huang, G., Xu, C., Guo, H., & Feng, Y. (2025). Photovoltaic cell surface defect detection via subtle defect enhancement and background suppression. Micromachines, 16(9), Article 1003. https://doi.org/10.3390/mi16091003
20. Tella, H., Hussein, A., Rehman, S., Liu, B., Balghonaim, A., & Mohandes, M. (2025). Solar photovoltaic panel cells defects classification using deep learning ensemble methods. Case Studies in Thermal Engineering, 66, Article 105749. https://doi.org/10.1016/j.csite.2025.105749
21. Thakfan, A., & Bin Salamah, Y. (2024). Artificial-intelligence-based detection of defects and faults in photovoltaic systems: A survey. Energies, 17(19), Article 4807. https://doi.org/10.3390/en17194807
22. Xiao, M., Yang, B., Wang, S., Zhang, Z., & He, Y. (2023). Fine coordinate attention for surface defect detection. Engineering Applications of Artificial Intelligence, 123, Article 106368. https://doi.org/10.1016/j.engappai.2023.106368
23. Zhang, Z., Yang, Y., & Jian, X. (2025). Defect analysis using electroluminescence imaging in photovoltaic modules. Measurement Science and Technology, 36(8), Article 085001. https://doi.org/10.1088/1361-6501/adf657