Infrared Defect Detection for Power System Photovoltaic Panels in UAV Inspection Using Dynamic Convolution
DOI:
https://doi.org/10.4108/ew.14715Keywords:
Dynamic Convolution, Infrared Image Detection, Lightweight Models, Multi-core Aggregation Strategy, Photovoltaic Power SystemsAbstract
INTRODUCTION: This paper addresses the significant demand for non-contact, efficient, and large-area inspections in photovoltaic power system drone infrared inspections. It also improves the performance bottlenecks of existing lightweight models in handling the low-resolution, high-noise, and complex features characteristic of real-world infrared images from power plants.
OBJECTIVES: The objective is to propose an Infrared Adaptive Dynamic Perceptron (IADP) that enhances defect recognition accuracy and hot spot detection precision for drone-based infrared inspections, while maintaining low computational complexity and achieving real-time performance on resource-constrained embedded platforms.
METHODS: The core method employs a dynamic convolution mechanism that aggregates multiple heterogeneous kernels based on input-dependent attention. The key innovations focus on the resource limitations of drones: a global average pooling method tailored to infrared images optimizes the attention mechanism, and the introduction of the Softplus activation function mitigates the loss of low-contrast features. Dynamic convolution kernels (DK-conv) are used to effectively adapt to the regular textures and localized defects of photovoltaic panels. A hierarchical temperature annealing training strategy is introduced to address training stability issues.
RESULTS: Experiments on a dataset of 11,608 drone infrared images show that the IADP-MobileNetV3-Small model improves defect recognition accuracy (TOP-1) by 4.0% and hot spot detection precision (mAP@0.5) by 4.7%, with only a 3.5% increase in computational complexity. It achieves 40.1 FPS on an embedded platform, exceeding the 20 FPS engineering requirement.
CONCLUSION: The proposed IADP model effectively meets the demands for real-time, accurate, and efficient infrared inspection in photovoltaic power plants using drones, offering a good balance between performance and computational cost for embedded deployment.
Downloads
References
[1] Alghanem H, Buckley A. Global benchmarking and modelling of installed solar photovoltaic capacity by country. Energies, 2024, 17(8): 1812.
[2] Li Y, Huang H, Xie Q, et al. Research on a surface defect detection algorithm based on MobileNet-SSD. Applied Sciences, 2018, 8(9): 1678.
[3] Liu G, Wen H. Printed circuit board defect detection based on MobileNet-Yolo-Fast. Journal of Electronic Imaging, 2021, 30(4): 043004-043004.
[4] Figgis B, Azid S I, Parlevliet D. Review of unmanned ground vehicles for PV plant inspection[J]. Solar Energy, 2025, 291: 113404.
[5] Quater P B, Grimaccia F, Leva S, et al. Light Unmanned Aerial Vehicles (UAVs) for cooperative inspection of PV plants[J]. IEEE Journal of Photovoltaics, 2014, 4(4): 1107-1113.
[6] Kirubakaran V, Preethi D M D, Arunachalam U, et al. Infrared thermal images of solar PV panels for fault identification using image processing technique[J]. International Journal of Photoenergy, 2022, 2022(1): 6427076.
[7] Ng H F. Automatic thresholding for defect detection. Pattern recognition letters, 2006, 27(14): 1644-1649.
[8] Daut D G, Zhao D. A flaw detection method based on morphological image processing. IEEE transactions on circuits and systems for video technology, 1993, 3(6): 389-398.
[9] Liu R, Lu Y, Gong C, et al. Infrared point target detection with improved template matching. Infrared Physics & Technology, 2012, 55(4): 380-387.
[10] Afifah A N N, Suyuti A. Hotspot detection in photovoltaic module using Otsu thresholding method [C]//2020 IEEE international conference on communication, networks and satellite (Comnetsat). IEEE, 2020: 408-412.
[11] Ou J, Wang J, Xue J, et al. Infrared image target detection of substation electrical equipment using an improved faster R-CNN. IEEE transactions on power delivery, 2023, 38(1): 387-396.
[12] Yang W, Li Y. Infrared detection and identification of rotating machinery based on faster R-CNN. Infrared Technology, 2020, 42(11): 1053-1060.
[13] Ma ZQ, Guo H, Wen H, et al. LFS-YOLO: A PV panel defect detection algorithm for drone infrared sensors. IEEE Sensors Journal. 2025;25(11):19592-19601.
[14] Tajwar T, Mobin O H, Khan F R, et al. Infrared thermography based hotspot detection of photovoltaic module using YOLO[C]//2021 IEEE 12th Energy Conversion Congress & Exposition-Asia (ECCE-Asia). IEEE, 2021: 1542-1547.
[15] Bing L I, Kuan Z, Yunshan B A I, et al. Defect detection of photovoltaic panel infrared image based on YOLOv7-EPAN. Infrared Technology, 2024, 46(11): 1315-1324.
[16] Liang R, Cen Y, Zhang L, et al. Innovative lightweight detection for airborne remote sensing: integrating G-shuffle and dynamic multiscale pyramid networks. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2025, 18: 2010-2023.
[17] Yang Y, Han J. Real-Time object detector based MobileNetV3 for UAV applications. Multimedia Tools and Applications, 2023, 82(12): 18709-18725.
[18] Zhang Y, Zhang J, Wang Q, et al. Dynet: Dynamic convolution for accelerating convolutional neural networks. arXiv preprint arXiv:2004.10694, 2020.
[19] Han Y, Huang G, Song S, et al. Dynamic neural networks: A survey. IEEE transactions on pattern analysis and machine intelligence, 2022, 44(11): 7436-7456.
[20] Jia Y, Zhao Q. Enhanced Photovoltaic–You Only Look Once Model for Photovoltaic Panel Fault Detection Based on IR Thermal Imaging Sensors. Sensors and Materials, 2026, 38(4): 1959-1974.
[21] Yang B, Bender G, Le Q V, et al. Condconv: Conditionally parameterized convolutions for efficient inference. Advances in neural information processing systems, 2019, 32.
[22] Chen Y, Dai X, Liu M, et al. Dynamic convolution: Attention over convolution kernels. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 2020:11027-11036.
[23] Li C, Zhou A, Yao A. Omni-dimensional dynamic convolution. arXiv preprint arXiv:2209.07947, 2022.
[24] Li X, Wang W, Hu X, et al. Selective kernel networks[C]//Proceedings of the IEEE/CVF conference on computer vision and pattern recognition. 2019: 510-519.
[25] Zhu X, Hu H, Lin S, Dai J. Deformable ConvNets v2: More deformable, better results. In: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). 2019:9300-9308.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Xiaoyang Hu, Donghui Xi, Chuande Hou, Shuhan Chen, Zehua Zhang
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
This is an open-access article distributed under the terms of the Creative Commons Attribution CC BY 4.0 license, which permits unlimited use, distribution, and reproduction in any medium so long as the original work is properly cited.
