基于改进YOLOv11n的光伏板异物与缺陷检测模型研究

doi:10.3778/j.issn.1002-8331.2412-0381

摘要/Abstract

摘要： 针对YOLOv11n在光伏板异物与缺陷检测中，检测精度低、计算资源要求高的问题，提出一种基于改进YOLOv11n的光伏板异物与缺陷检测算法FESI-YOLOv11n。采用C3k2_Faster_EMA模块替换C3k2模块，扩展初始卷积通道数，使网络能更高效地进行多尺度特征提取；提出一种重构的检测头，将多分支、多尺度思想与重参数化思想结合，提高单一卷积的特征提取能力；在特征融合前增加注意力机制模块SEAttention，减少计算量；使用Inner_DIoU损失函数代替CIoU损失函数，弥补边界框回归方法的不足，进一步提高检测能力。实验结果表明，与YOLOv11n模型相比，改进后的算法mAP50提高了3.6个百分点，mAP50-95提高了3.4个百分点，模型的参数量降低了21.29%，计算量降低了25.4%，证明改进后的算法能够更好地应用在光伏板异物与缺陷检测的任务中。

关键词: 光伏板, YOLOv11n, 异物检测, 缺陷检测

Abstract: To address the issues of low detection accuracy and high computational resource demands of YOLOv11n in photovoltaic panel foreign object and defect detection, this paper proposes an improved YOLOv11n-based algorithm named FESI-YOLOv11n. Firstly, the C3k2 module is replaced with the C3k2_Faster_EMA module, which expands the initial convolutional channels to enable more efficient multi-scale feature extraction. Secondly, a reconstructed detection head is proposed, integrating multi-branch and multi-scale design with re-parameterization to enhance the feature extraction capability of single convolutions. Thirdly, an SEAttention mechanism is added before feature fusion to reduce computational overhead. Finally, the Inner_DIoU loss function replaces the CIoU loss to address limitations in bounding box regression and further improve detection performance. Experimental results demonstrate that compared to the original YOLOv11n model, the improved algorithm achieves a 3.6 percentage points increase in mAP50 and a 3.4 percentage points improvement in mAP50-95, while reducing model parameters by 21.29% and computational load by 25.4%. These advancements validate the superior applicability of the proposed algorithm in detecting foreign objects and defects in photovoltaic panels.

Key words: solar photovoltaic panels, YOLOv11n, foreign object detection, defect detection

韩涛, 于帅帅, 马玲, 黄友锐, 侯帅男, 庞家乐. 基于改进YOLOv11n的光伏板异物与缺陷检测模型研究[J]. 计算机工程与应用, 2025, 61(14): 123-134.

HAN Tao, YU Shuaishuai, MA Ling, HUANG Yourui, HOU Shuainan, PANG Jiale. Research on Detection Model of Foreign Objects and Defects in Photovoltaic Panels Based on Improved YOLOv11n[J]. Computer Engineering and Applications, 2025, 61(14): 123-134.

参考文献

[1] 肖佳, 梅琦, 黄晓琪, 等. “双碳”目标下我国光伏发电技术现状与发展趋势[J]. 天然气技术与经济, 2022, 16(5): 64-69.
XIAO J, MEI Q, HUANG X Q, et al. Status quo and development trend of photovoltaic power-generating technology under the dual-carbon goal[J]. Natural Gas Technology and Economy, 2022, 16(5): 64-69.
[2] 李明东, 李婧雯. “双碳” 目标下中国分布式光伏发电的发展现状和展望[J]. 太阳能, 2023(5): 5-10.
LI M D, LI J W. Development status and prospects of distributed PV power generation in China under the goal of emission peak and carbon neutrality[J]. Solar Energy, 2023(5): 5-10.
[3] 吴涛, 赖菲. 基于LeNet-5模型的太阳能电池板缺陷识别分类[J]. 热力发电, 2019, 48(3): 120-125.
WU T, LAI F. Identification and classification of defects in solar cells based on LeNet-5 model[J]. Thermal Power Generation, 2019, 48(3): 120-125.
[4] 赵强, 刘胜杰, 韩东成, 等. 基于改进K均值聚类的光伏板缺陷检测方法[J]. 红外技术, 2024, 46(4): 475-482.
ZHAO Q, LIU S J, HAN D C, et al. Photovoltaic panel defect detection method based on improved K-means clustering[J]. China Industrial Economics, 2024, 46(4): 475-482.
[5] 管宽岐, 蔺雨桐, 赵雨薇, 等. 基于深度学习的航拍光伏板红外图像热斑检测方法研究[J]. 电子测量技术, 2022, 45(22): 75-81.
GUAN K Q, LIN Y T, ZHAO Y W, et al. Photovoltaic hot spot detection of aerial infrared image based on deep learning[J]. Electronic Measurement Technology, 2022, 45(22): 75-81.
[6] 张猛, 尹丽菊, 周辉, 等. 基于SimAM-Ada YOLOv5的太阳能电池表面缺陷检测[J]. 电子测量技术, 2023, 46(22): 17-25.
ZHANG M, YIN L J, ZHOU H, et al. Surface defect detection of solar cells based on SimAM-Ada YOLOv5[J]. Electronic Measurement Technology, 2023, 46(22): 17-25.
[7] YANG L X, ZHANG R Y, LI L D, et al. SimAM: a simple, parameter-free attention module for convolutional neural networks[C]//Proceedings of the 38th International Conference on Machine Learning, 2021: 11863-11874.
[8] LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot MultiBox detector[C]//Proceedings of the European Conference on Computer Vision. Cham: Springer International Publishing, 2016: 21-37.
[9] REN S, HE K, GIRSHICK R, et al. Faster R-CNN: towards real-time object detection with region proposal networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149.
[10] 何湘杰, 宋晓宁. YOLOv4-Tiny的改进轻量级目标检测算法[J]. 计算机科学与探索, 2024, 18(1): 138-150.
HE X J, SONG X N. Improved YOLOv4-tiny lightweight target detection algorithm[J]. Journal of Frontiers of Computer Science and Technology, 2024, 18(1): 138-150.
[11] 王春梅, 刘欢. YOLOv8-VSC: 一种轻量级的带钢表面缺陷检测算法[J]. 计算机科学与探索, 2024, 18(1): 151-160.
WANG C M, LUI H. YOLOv8-VSC: a lightweight surface defect detection algorithm for strip steel[J]. Journal of Frontiers of Computer Science and Technology, 2024, 18(1): 151-160.
[12] 李峻宇, 刘乾坤, 付莹. 融合注意力机制的红外小目标检测[J]. 航空学报, 2024, 45(14): 90-101.
LI J Y, LIU Q K, FU Y. Infrared small target detection based on attention mechanism[J]. Acta Aeronautica ET Astronautica Sinica, 2024, 45(14): 90-101.
[13] 赵晓雨, 高林, 杨校李, 等. 融合注意力机制的YOLOv5光伏板电致发光图像缺陷检测算法[J]. 湖北民族大学学报 (自然科学版), 2023, 41(1): 65-70.
ZHAO X Y, GAO L, YANG X L, et al. Algorithm of YOLOv5 fusing attention mechanism for defect detection of photovoltaic panel electroluminescent image[J]. Journal of Hubei Minzu University (Natural Science Edition), 2023, 41(1): 65-70.
[14] WANG Q L, WU B G, ZHU P F, et al. ECA-Net: efficient channel attention for deep convolutional neural networks[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 11531-11539.
[15] WU W, LIU H, LI L, et al. Application of local fully convolutional neural network combined with YOLOv5 algorithm in small target detection of remote sensing image[J]. PLoS One, 2021, 16(10): e0259283.
[16] PAN Y Z, REN C, WU X H, et al. Real image denoising via guided residual estimation and noise correction[J]. IEEE Transactions on Circuits and Systems for Video Technology, 2023, 33(4): 1994-2000.
[17] TERVEN J, CóRDOVA-ESPARZA D M, ROMERO-GONZáLEZ J A. A comprehensive review of YOLO architectures in computer vision: from YOLOv1 to YOLOv8 and YOLO-NAS[J]. Machine Learning and Knowledge Extraction, 2023, 5(4): 1680-1716.
[18] 钟泳松, 徐凌桦, 周克. 基于改进SSD算法的光伏组件缺陷检测研究[J]. 微处理机, 2022, 43(1): 22-25.
ZHONG Y S, XU L H, ZHOU K. Research on defect detection of photovoltaic module based on improved SSD algorithm[J]. Microprocessors, 2022, 43(1): 22-25.
[19] ALI F, REDMON J. YOLOv3: an incremental improvement[J]. arXiv:1804.02767, 2018.
[20] GUO S Q, GUO Z W, WANG Y L, et al. Detection method of photovoltaic panel defect based on improved mask R-CNN[J]. Journal of Internet Technology, 2022, 23(2): 397-406.
[21] REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified, real-time object detection[C]//Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2016: 779-788.
[22] JEGHAM N, CHAN Y K, MARWAN A, et al. Evaluating the evolution of YOLO (you only look once) models: a comprehensive benchmark study of YOLO11 and its predecessors[J]. arXiv:2411.00201, 2024.
[23] FAKHOURI F, FAKHOURI F, FRéMEAUX-BACCHI V, et al. C3 glomerulopathy: a new classification[J]. Nature Reviews Nephrology, 2010, 6(8): 494-499.
[24] HU J, LL S, GANG S, et al. Squeeze-and-excitation networks[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2018: 7132-7141.
[25] TAN M X, PANG R M, LE Q V. EfficientDet: scalable and efficient object detection[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway: IEEE, 2020: 10778-10787.
[26] ZHANG H, XU C, ZHANG S J. Inner-IoU: more effective intersection over union loss with auxiliary bounding box[J]. arXiv:2311.02877, 2023.
[27] ZHANG X, LIU C, YANG D G, et al. RFAConv: innovating spatial attention and standard convolutional operation[J]. arXiv:2304.03198, 2023.
[28] ZAK I, KATZ R, KLEIN I. MLCA-a machine learning framework for INS coarse alignment[J]. Sensors (Basel), 2020, 20(23): 6959.
[29] ZHOU H Y, LI J X, PENG J Q, et al. Triplet attention: rethinking the similarity in transformers[C]//Proceedings of the 27th ACM SIGKDD Conference on Knowledge Discovery & Data Mining. New York: ACM, 2021: 2378-2388.
[30] YANG Z M, WANG X L, LI J G. EIoU: an improved vehicle detection algorithm based on VehicleNet neural network[J]. Journal of Physics: Conference Series, 2021, 1924(1): 012001.
[31] BIST R B, SUBEDI S, YANG X, et al. A novel YOLOv6 object detector for monitoring piling behavior of cage-free laying hens[J]. Agriengineering, 2023, 5(2): 905-923.
[32] ALSHAHRANI A, ALMATRAFI M M, MUSTAFA J I, et al. A children’s psychological and mental health detection model by drawing analysis based on computer vision and deep learning[J]. Engineering, Technology & Applied Science Research, 2024, 14(4): 15533-15540.
[33] WANG A, CHEN H, LIU L H, et al. YOLOv10: real-time end-to-end object detection[J]. arXiv:2405.14458, 2024.