https://doi.org/10.65770/CKUL9343
ABSTRACT
This paper deals with the challenge of classifying solar panel images as clean or dusty, at a scale that matters to reduce photovoltaic-related costs and inspection efforts. To solve this challenge in a transparent and computationally efficient way, we propose a novel residual depth-wise separable lightweight Inception-based model created from scratch. The backbone is based on bottleneck (1×1) convolutions and spatially separable (1×3) and (3×1) convolutions grouped in multi-branch Inception style residual blocks, and is also informed by a convolution only attention module that merges bottleneck and spatially separable convolutions to produce feature re-weighting maps without utilizing pre-existing attention frameworks such as SE or CBAM. The model is learned on realistic solar panel dust data with disjoint training, validation, and testing sets. The images are rescaled at a fixed resolution and augmented through geometric transformations so as to enhance generalization capabilities. Quantitative experiments test the trained network on standard classification metrics, with the addition of confusion matrices and precision-recall curves to observe the performance of each class in detail, including accuracy, precision, recall, F1-score and ROC–AUC. Grad-CAM and LIME are embedded as XAI tools to show which areas of the solar panel images lead the model to its predictions. The results, in general, suggest that the lightweight and interpretable network design is capable of capturing discriminative dust related patterns, while the XAI analysis discloses the potential and existing limitations, as well as clear directions on the further enhancement of the methodology.
References
- [1] F. Shaaban, A. Alarif, M. Mokhtar, U. Tariq, A. H. Osman, and A. R. Al-Ali, ‘A New Data-Based Dust Estimation Unit for PV Panels,” Energies, vol. 13, no. 14, p. 3601, July 2020, doi: 10.3390/en13143601.
- [2] Mokhtar and M. F. Shaaban, ‘A New ANN-Based Cleaning Approach for Photovoltaic Solar Panels’, in 2022 9th International Conference on Electrical and Electronics Engineering (ICEEE), Alanya, Turkey: IEEE, Mar. 2022, pp. 260–263. doi: 10.1109/ICEEE55327.2022.9772579.
- [3] Ö. F. Alçin, M. Aslan, and A. Ari, ‘SolPowNet: Dust Detection on Photovoltaic Panels Using Convolutional Neural Networks’, Electronics, vol. 14, no. 21, p. 4230, Oct. 2025, doi: 10.3390/electronics14214230.
- [4] M. Alatwi, H. Albalawi, A. Wadood, H. Anwar, and H. M. El-Hageen, ‘Deep Learning-Based Dust Detection on Solar Panels: A Low-Cost Sustainable Solution for Increased Solar Power Generation’, Sustainability, vol. 16, no. 19, p. 8664, Oct. 2024, doi: 10.3390/su16198664.
- [5] Gao and S. Li, ‘A deep learning-based method detects dust from solar PV panels through Unmanned Aerial Vehicles’, J. Phys.: Conf. Ser., vol. 2584, no. 1, p. 012019, Sept. 2023, doi: 10.1088/1742-6596/2584/1/012019.
- [6] Naeem, K. Chadda, S. Vahaji, J. Ahmad, X. Li, and E. Asadi, ‘Aerial Imaging-Based Soiling Detection System for Solar Photovoltaic Panel Cleanliness Inspection’, Sensors, vol. 25, no. 3, p. 738, Jan. 2025, doi: 10.3390/s25030738.
- [7] Fang, P. M. Rodrigo, L. R. Narvarte, G. Makrides, and G. E. Georghiou, “A deep learning model for photovoltaic soiling loss estimation,” Applied Energy, vol. 376, pt. B, art. no. 124335, Dec. 2024, doi: 10.1016/j.apenergy.2024.124335.
- [8] Cavieres, A. Barraza, A. Estay, M. Bilbao, and A. Valdivia-Lefort, “Automatic soiling and partial shading assessment tool for photovoltaic power plants based on convolutional neural networks,” Applied Energy, vol. 306, art. no. 117964, Jan. 2022, doi: 10.1016/j.apenergy.2021.117964.
- [9] Zhang, Y. Zhang, S. Chen, D. Zhou, M. T. M. Emam, and L. Yu, “SoilingEdge: PV soiling power loss estimation at the edge using surveillance cameras,” IEEE Transactions on Sustainable Energy, vol. 15, no. 1, pp. 556–566, Jan. 2024, doi: 10.1109/TSTE.2023.3320690.
- Naskar, A. Chowdhury, S. Adhikari, and C. Chakraborty, “A deep learning regression framework for robust PV soiling quantification using multi-source drone-ground imaging,” Expert Systems with Applications, vol. 306, art. no. 130944, Apr. 2026, doi: 10.1016/j.eswa.2025.130944.
- Micheli, E. F. Fernández, A. Smets, and J. A. del Cañizo, “Variability and associated uncertainty analysis of image-based soiling characterization,” Solar Energy Materials and Solar Cells, vol. 259, art. no. 112437, Jun. 2023, doi: 10.1016/j.solmat.2023.112437.
- Yang, W. Javed, B. Guo, and J. Ji, “Estimating PV soiling loss using panel images and a feature-based regression model,” IEEE Journal of Photovoltaics, vol. 14, pp. 661–668, 2024, doi: 10.1109/JPHOTOV.2024.3388168.
Download all article in PDF
