基于改进YOLO11n的光伏板缺陷检测方法

doi:10.16088/j.issn.1001-6600.2025071102

摘要/Abstract

摘要： 针对光伏板缺陷检测算法中远景小目标缺陷特征易被弱化及模型复杂度高等问题,本文提出一种改进YOLO11n的轻量化光伏板缺陷检测算法(FEM-YOLO)。首先,融合FasterBlock和EMA来改进C3k2模块,构建C3k2-Faster-EMA结构,以增强网络对缺陷目标特征的学习与捕捉能力。其次,在C2PSA中加入Mona模块,优化模型的特征提取与表达能力。此外,在主干网络加入注意力机制MLCA,提升多样化目标的特征提取鲁棒性。最后,增加P2小目标检测层并设计高效检测头EfficientHead,提升对微小缺陷的捕获能力,同时降低模型复杂程度。实验结果表明,与YOLO11n模型相比,改进后的算法mAP50和mAP50-95都提升1.9个百分点,模型参数量降至2.1×10⁶,存储体积压缩至4.4 MiB,改进算法在保证检测精度提升的同时大幅降低了模型复杂度。

关键词: 光伏板, 缺陷检测, 轻量化模型, 小目标检测, YOLO11n

Abstract: To address the issues of weakened features for distant small-target defects and high model complexity in photovoltaic panel defect detection algorithms, this study proposes an improved lightweight algorithm named FEM-YOLO. Firstly, the C3k2 module is enhanced by integrating FasterBlock and EMA, constructing a C3k2-Faster-EMA structure to improve the network's ability to learn and capture features of defective targets. Subsequently, the Mona module is incorporated into the C2PSA block, optimizing the model's feature extraction and representation capabilities. Moreover, the MLCA mechanism is integrated into the backbone network to enhance the robustness of feature extraction for diverse targets. Finally, an additional P2 detection layer is added specifically for small targets, and an efficient detection head named EfficientHead is designed. This combination enhances the capability to capture micro-defects while simultaneously reducing model complexity. Experimental results demonstrate that, compared with the original YOLO11n model, the improved algorithm achieves increases of 1.9% in both mAP50 and mAP50-95 metrics. Furthermore, the model parameter count is reduced to 2.1×10⁶ and the model size is compressed to 4.4 MiB. Thus, the proposed FEM-YOLO algorithm significantly enhances detection accuracy while substantially reducing model complexity.

Key words: photovoltaic panel, defect detection, lightweight model, small object detection, YOLO11n

中图分类号: TP391.41

杨云波, 南新元, 蔡鑫. 基于改进YOLO11n的光伏板缺陷检测方法[J]. 广西师范大学学报（自然科学版）, 2026, 44(3): 47-59.

YANG Yunbo, NAN Xinyuan, CAI Xin. Photovoltaic Panel Defect Detection Method Based on Improved YOLO11n[J]. Journal of Guangxi Normal University(Natural Science Edition), 2026, 44(3): 47-59.

参考文献

[1] 钟洪麟, 李丁丁, 刘茜, 等. 全球退役光伏组件回收研究热点、演化趋势与展望[J]. 生态学报, 2025, 45(9): 4079-4106. DOI: 10.20103/j.stxb.202410232588.
[2] NGUYEN S D, SONG J H, TRAN V P, et al. DEPP: Automated detection of pavement patching and nonslip coatings[J]. Measurement, 2025, 252: 117315. DOI: 10.1016/j.measurement.2025.117315.
[3] WEI W, CHENG Y, HE J F, et al. A review of small object detection based on deep learning[J]. Neural Computing and Applications, 2024, 36(12): 6283-6303. DOI: 10.1007/s00521-024-09422-6.
[4] MA Y X, YIN J X, HUANG F, et al. Surface defect inspection of industrial products with object detection deep networks: a systematic review[J]. Artificial Intelligence Review, 2024, 57(12): 333. DOI: 10.1007/s10462-024-10956-3.
[5] 鲁东林, 王淑青, 鲁濠, 等. 一种改进Faster R-CNN的太阳能电池片缺陷检测方法[J]. 激光杂志, 2022, 43(3): 50-55. DOI: 10.14016/j.cnki.jgzz.2022.03.050.
[6] 焦思韬, 王可庆, 周奇, 等. 基于改进SSD算法的光伏板缺陷检测研究[J]. 软件, 2023, 44(12): 47-52.
[7] 万鸿炜, 陈平华. 基于Involution算子和协调反向注意力的息肉图像分割[J]. 计算机与现代化, 2024(11): 84-90, 98.
[8] HUANG C, PANG Z, XU J Z. Detection method of manipulator grasp pose based on RGB-D image[J]. Neural Processing Letters, 2024, 56(4): 211. DOI: 10.1007/s11063-024-11662-5.
[9] 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. DOI: 10.19651/j.cnki.emt.2313181.
[10] XUE Y H, WANG J H, HU S T, et al. Persistent Calyx of Rosa roxburghii recognition based on SimAM-YOLO-v5s[J].Journal of Food Measurement and Characterization, 2025, 19(7): 4715-4726. DOI: 10.1007/s11694-025-03284-9.
[11] 朱成杰, 刘乐乐, 朱洪波. 多尺度的YOLOv8-MNS光伏板缺陷检测算法[J/OL]. 重庆工商大学学报(自然科学版), 2024: 1-9. (2024-06-04). https://kns.cnki.net/KCMS/detail/detail.aspx?filename=YZZK20240531002&dbname=CJFD&d
bcode=CJFQ.
[12] LI J H, HU T, LIAN X Y, et al. Constitutive optimization modeling of magnetorheological dampers under multiple influencing factors[J]. International Journal of Mechanical Sciences, 2025, 295: 110284. DOI: 10.1016/j.ijmecsci.2025.110284.
[13] 苏俊, 唐潮龙, 刘智权, 等. 基于全局与局部特征提取增强的光伏板缺陷检测算法[J]. 计算机工程与应用, 2025, 61(12): 299-310.
[14] HUANG Y Y, WANG D, WU B X, et al. NST-YOLO11: ViT merged model with neuron attention for arbitrary-oriented ship detection in SAR images[J]. Remote Sensing, 2024, 16(24): 4760. DOI: 10.3390/rs16244760.
[15] HE L H, ZHOU Y Z, LIU L, et al. Research on the directional bounding box algorithm of YOLO11 in tailings pond identification[J]. Measurement, 2025, 253: 117674. DOI: 10.1016/j.measurement.2025.117674.
[16] ZHAO Y Z, ZHAO H D, SHI J F. YOLOv8-MAH: multi-attribute recognition model for vehicles[J]. Pattern Recognition, 2025, 167: 111849. DOI: 10.1016/j.patcog.2025.111849.
[17] ZHONG H, ZHANG Y, SHI Z G, et al. PS-YOLO: a lighter and faster network for UAV object detection[J]. Remote Sensing, 2025, 17(9): 1641. DOI: 10.3390/rs17091641.
[18] LIU X H, CHEN D F, WANG X B, et al. Rep-MCA-former: an efficient multi-scale convolution attention encoder for text-independent speaker verification[J]. Computer Speech & Language, 2024, 85: 101600. DOI: 10.1016/j.csl.2023.101600.
[19] ZHAO Q, ZHU J H. An Improved YOLOv11 architecture with multi-scale attention and spatial fusion for fine-grained residual detection[J]. Results in Engineering, 2025, 27: 107061. DOI: 10.1016/j.rineng.2025.107061.
[20] ZHANG M J, HONG D, WU J B, et al. Sheep-YOLO: improved and lightweight YOLOv8n for precise and intelligent recognition of fattening lambs' behaviors and vitality statuses[J]. Computers and Electronics in Agriculture, 2025, 236: 110413. DOI: 10.1016/j.compag.2025.110413.
[21] 陈星. 网联环境下车辆行驶特征提取及数据可信甄别研究[D]. 北京: 中国人民公安大学, 2025. DOI: 10.27634/d.cnki.gzrgu.2025.000331.
[22] LUO Y, LING J, WANG J W, et al. SFW-YOLO: a lightweight multi-scale dynamic attention network for weld defect detection in steel bridge inspection[J]. Measurement, 2025, 253: 117608. DOI: 10.1016/j.measurement.2025.117608.
[23] JIN S X, ZHOU L, ZHOU H P. CO-YOLO: a lightweight and efficient model for Camellia oleifera fruit object detection and posture determination[J]. Computers and Electronics in Agriculture, 2025, 235: 110394. DOI: 10.1016/j.compag.2025.
110394.
[24] 何珠. 基于卷积神经网络的航空图像目标检测研究与应用[D]. 乌鲁木齐: 新疆大学, 2022. DOI: 10.27429/d.cnki.gxjdu.2022.001714.
[25] 吕辉, 司可. 基于改进RT-DETR的光伏板缺陷检测[J]. 广西师范大学学报(自然科学版), 2026, 44(2): 52-64. DOI: 10.16088/j.issn.1001-6600.2025060302.
[26] SHETTY R, AL MAJALI A, WELLS L. Predicting surface roughness in foundry applications through MPCA and convolutional neural networks[J]. International Journal of Metalcasting, 2025. DOI: 10.1007/s40962-025-01618-3.
[27] 彭朋, 高浪超, 李家春. 基于改进Yolov8-GCB的公路落石检测方法[J]. 长安大学学报(自然科学版), 2025, 45(2): 24-35. DOI: 10.19721/j.cnki.1671-8879.2025.02.003.
[28] LI Y B, ZHANG W W, LV S T, et al. YOLOv11-CAFM model in ground penetrating radar image for pavement distress detection and optimization study[J]. Construction and Building Materials, 2025, 485: 141907. DOI: 10.1016/j.conbuildmat.
2025.141907.
[29] YE J G, SHU Z L, ZHOU W, et al. YOLOv8 architectural scene section recognition method based on SimAM-EMA hybrid attention mechanism[J]. Sensors, 2025, 25(10): 3060. DOI: 10.3390/s25103060.
[30] HACHAJ T, PIEKARCZYK M. On explainability of reinforcement learning-based machine learning agents trained with proximal policy optimization that utilizes visual sensor data[J]. Applied Sciences, 2025, 15(2): 538. DOI: 10.3390/
app15020538.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed