Deteksi Kerusakan Modul Surya Menggunakan Faster R-CNN ResNet-50

Fathirul Ikhsan; Rizqia Cahyaningtyas; Dwina Kuswardani

doi:10.35760/tr.2025.v30i3.116

Authors

Fathirul Ikhsan Institut Teknologi Perusahaan Listrik Negara Author
Rizqia Cahyaningtyas Institut Teknologi Perusahaan Listrik Negara Author
Dwina Kuswardani Institut Teknologi Perusahaan Listrik Negara Author

DOI:

https://doi.org/10.35760/tr.2025.v30i3.116

Keywords:

Faster R-CNN, object detection, ResNet-50, RGB image, solar module defect detection

Abstract

Solar modules play a crucial role in photovoltaic power generation systems, yet their performance can degrade due to physical and electrical damage. Therefore, automatic inspection is required to improve maintenance efficiency and prevent long-term performance loss. This study aims to implement an object detection approach for identifying solar module defects from visible RGB images using Faster R-CNN with a ResNet-50 backbone. The dataset was obtained from the Kaggle platform and manually annotated into PASCAL VOC format with two defect classes, namely physical damage and electrical damage, and expanded through data augmentation. The model was trained under several training configurations and evaluated using mean Average Precision (mAP), precision, recall, F1-score, and accuracy. The best performance was achieved using a batch size of 8, learning rate of 0.0001, and 30 epochs, resulting in 89% accuracy and 93% mAP. The results indicate that the model consistently detects both defect types and demonstrates potential for automated solar module inspection.

Downloads

Download data is not yet available.

References

[1] J. Sundah et al., “Solar power generation system at household scale,” Interdiscip. J. Adv. Res. Innov., vol. 2, no. 4, pp. 79–82, 2023, doi: 10.58860/ijari.v2i4.57.

[2] G. Wang, Z. Zhang, and J. Lin, “Multi-energy complementary power systems based on solar energy: A review,” Renew. Sustain. Energy Rev., vol. 199, p. 114464, Jul. 2024, doi: 10.1016/j.rser.2024.114464.

[3] S. Gyamfi, B. Aboagye, M. Obeng, and F. Peprah, “Criticality and severity of adverse effects of the sun on performance of solar PV systems,” Solar Energy Adv., vol. 4, p. 100058, 2024, doi: 10.1016/j.seja.2024.100058.

[4] S. Fuad, I. Setiawan, and T. Andromeda, “Optimasi algoritma firefly pada maximum power point tracking (MPPT) saat kondisi panel surya terhalangi sebagian,” Multitek Indones. J. Ilm., vol. 16, no. 1, pp. 21–36, Jul. 2022, doi: 10.24269/mtkind.v16i1.4844.

[5] M. E. Emetere and S. A. Afolalu, “Challenges of small-scale standalone solar energy supply in parts of Lagos via 39 years dataset,” E3S Web Conf., vol. 294, p. 01007, 2021, doi: 10.1051/e3sconf/202129401007.

[6] A. M. Gabor, P. J. Knodle, M. Covino, J. Fioresi, D. J. Colvin, and K. O. Davis, “Electroluminescence analysis and grading of hail damaged solar panels,” IEEE J. Photovolt., vol. 12, no. 6, pp. 1308–1313, Nov. 2022, doi: 10.1109/JPHOTOV.2022.3206906.

[7] J. A. Dhanraj et al., “An effective evaluation on fault detection in solar panels,” Energies, vol. 14, no. 22, p. 7770, Nov. 2021, doi: 10.3390/en14227770.

[8] E. K. Ukiwe, S. A. Adeshina, T. Jacob, and A. M. Zungeru, “Deep learning model for detection of hotspots using infrared thermographic images of electrical installations,” J. Electr. Syst. Inf. Technol., vol. 11, p. 24, Jun. 2024, doi: 10.1186/s43067-024-00148-y.

[9] R. Cahyaningtyas, S. Madenda, Bertalya, and D. Indarti, “Solar module defects classification using deep convolutional neural network,” Int. J. Adv. Intell. Inform, vol. 11, no. 3, pp. 499–513, Aug. 2025, doi: 10.26555/ijain.v11i3.1818.

[10] A. Shaik, A. Balasundaram, L. S. Kakarla, and N. Murugan, “Deep learning-based detection and segmentation of damage in solar panels,” Automation, vol. 5, no. 2, pp. 128–150, May 2024, doi: 10.3390/automation5020009.

[11] C. Bu, T. Liu, T. Wang, H. Zhang, and S. Sfarra, “A CNN-architecture-based photovoltaic cell fault classification method using thermographic images,” Energies, vol. 16, no. 9, p. 3749, Apr. 2023, doi: 10.3390/en16093749.

[12] S. Hassan and M. Dhimish, “A survey of CNN-based approaches for crack detection in solar PV modules: Current trends and future directions,” Solar, vol. 3, no. 4, pp. 663–683, Dec. 2023, doi: 10.3390/solar3040036.

[13] S. P. Pathak, S. Patil, and S. Patel, “Solar panel hotspot localization and fault classification using deep learning approach,” Procedia Comput. Sci., vol. 204, pp. 698–705, 2022, doi: 10.1016/j.procs.2022.08.084.

[14] X. Zhang, Y. Hao, H. Shangguan, P. Zhang, and A. Wang, “Detection of surface defects on solar cells by fusing multi-channel convolution neural networks,” Infrared Phys. Technol., vol. 108, p. 103334, Aug. 2020, doi: 10.1016/j.infrared.2020.103334.

[15] P. Ma'rifah, M. Sarosa, and R. Erfan, “Comparison of Faster R-CNN ResNet-50 and ResNet-101 methods for recycling waste detection,” Int. J. Comput. Appl. Technol. Res., vol. 12, no. 12, pp. 26–32, Dec. 2023, doi: 10.7753/IJCATR1212.1006.

[16] M. A. M. Amin, N. A. A. Jamaludin, and N. F. Awang, “Object detection model based on ResNet-50 within the effective R-CNN framework,” J. Inf. Syst. Eng. Manag., vol. 10, pp. 360–369, 2025, doi: 10.52783/jisem.v10i29s.4487.

[17] N. Bilous, V. Malko, M. Frohme, and A. Nechyporenko, “Comparison of CNN-based architectures for detection of different object classes,” AI, vol. 5, no. 4, pp. 2300–2320, Dec. 2024, doi: 10.3390/ai5040113.

[18] X. Renjun, Y. Junliang, W. Yi, and S. Mengcheng, “Fault detection method based on improved Faster R-CNN: Take ResNet-50 as an example,” Geofluids, vol. 2022, p. 7812410, 2022, doi: 10.1155/2022/7812410.

[19] Afroz, “Solar panel images,” Kaggle, 2023. [Online]. Available: https://www.kaggle.com/datasets/pythonafroz/solar-panel-images. (Accessed: 22-May-2025).

[20] Z. Luo, S. Y. Cheng, and Q. Y. Zheng, “GAN-based augmentation for improving CNN performance of classification of defective photovoltaic module cells in electroluminescence images,” in IOP Conf. Ser.: Earth Environ. Sci., vol. 354, no. 1, 2019, Art. no. 012106, doi: 10.1088/1755-1315/354/1/012106.

[21] N. A. M. Roslan, N. M. Diah, Z. Ibrahim, Y. Munarko, and A. E. Minarno, “Automatic plant recognition using convolutional neural network on Malaysian medicinal herbs: The value of data augmentation,” Int. J. Adv. Intell. Inform., vol. 9, no. 1, pp. 136–147, Mar. 2023, doi: 10.26555/ijain.v9i1.1076.

[22] S. Hassan and M. Dhimish, “Dual spin max pooling convolutional neural network for solar cell crack detection,” Sci. Rep., vol. 13, no. 1, Dec. 2023, doi: 10.1038/s41598-023-38177-8.

[23] A. Ratnakar, “Evaluating speech analysis techniques for Parkinson’s disease detection: A comparison of machine learning and deep learning algorithms,” Int. J. Adv. Res., vol. 12, no. 5, pp. 1118–1137, May 2024, doi: 10.21474/IJAR01/18827.

[24] D. P. Lestari et al., Komputer Vision, Kecerdasan Artifisial, dan Sistem Tertanam. Depok, Indonesia: Penerbit Gunadarma, 2024.

[25] L. Hakim, A. Hendrawan, and R. Khoiriyah, “Traffic vehicle detection using Faster R-CNN: A comparative analysis of backbone architectures,” Int. J. Artif. Intell. Sci., vol. 1, no. 1, pp. 50–62, Sep. 2024, doi: 10.63158/ijais.v1.i1.5.

[26] T. T. Habe, K. Haataja, and P. Toivanen, “Efficiency meets accuracy: Benchmarking object detection models for pathology detection in wireless capsule endoscopy,” IEEE Access, vol. 12, pp. 126793–126817, 2024, doi: 10.1109/ACCESS.2024.3456100.