针对现有光伏组件电致发光（EL）的缺陷检测中存在的背景干扰和计算冗余问题,以及模型精度与速度难以平衡的挑战,提出一种改进的YOLOv8光伏EL缺陷检测方法：YOLOv8-LSB。首先,在主干网络中引入SCConv卷积模块,以降低空间冗余并增强小目标特征提取能力;其次,在颈部添加LSK注意力机制,降低背景干扰;同时采用BiFPN结构提升多尺度特征融合能力,更好地捕捉不同方面的特征。最后,将Inner-CIoU作为边界框回归损失函数,提高回归精度和收敛速度。实验结果显示,YOLOv8-LSB在mAP@0.5上达91.2 %,FPS达170.2 帧/s,相较于基准模型YOLOv8n,平均精度提高2.6个百分点,FPS提升4.8 帧/s,实现了更实时且准确的光伏EL缺陷检测。
Abstract
This paper addresses the challenges of background interference, computational redundancy, and the difficulty in balancing model accuracy with processing speed in existing photovoltaic module electroluminescence （EL） defect detection methods. We propose an enhanced YOLOv8-based defect detection method, named YOLOv8-LSB, to tackle these issues. Firstly, we introduce the SCConv convolution module into the backbone network, reducing spatial redundancy while improving the extraction of small target features. Secondly, we incorporate the LSK attention mechanism in the neck to effectively mitigate background interference. Additionally, we use the BiFPN structure to enhance multi-scale feature fusion, enabling the model to capture features from various perspectives more effectively. Lastly, we employ Inner-CIoU as the bounding box regression loss function, which improves both regression accuracy and convergence speed. Experimental results show that YOLOv8-LSB achieves 91.2% mAP@0.5 and 170.2 FPS. Compared to the baseline model YOLOv8n, the proposed method improves average accuracy by 2.6 percentage points and FPS by 4.8 frames per second, making it a more real-time and accurate solution for photovoltaic EL defect detection.
关键词
目标检测 /
光伏组件 /
YOLOv8 /
注意力机制
Key words
object detection /
photovoltaic modules /
YOLOv8 /
attention mechanism
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] 邱修林, 宋博, 殷俊, 等. 基于强化学习的无人机智能组网技术及应用综述[J]. 哈尔滨工程大学学报, 2024, 45(8): 1576-1589, 1598.
QIU X L, SONG B, YIN J, et al.Review of unmanned aerial vehicle intelligent networking technology and applications based on reinforcement learning[J]. Journal of Harbin Engineering University, 2024, 45(8): 1576-1589, 1598.
[2] 何兴, 王旭, 许野, 等. 光伏产业与环境支撑体系的耦合协调研究[J]. 太阳能学报, 2023, 44(9): 194-203.
HE X, WANG X, XU Y, et al.Study on coupling coordination between photovoltaic industry and environmental support system[J]. Acta energiae solaris sinica, 2023, 44(9): 194-203.
[3] DALAL N, TRIGGS B.Histograms of oriented gradients for human detection[C]//2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'05). San Diego, CA, USA, 2005: 886-893.
[4] FELZENSZWALB P, MCALLESTER D, RAMANAN D.A discriminatively trained, multiscale, deformable part model[C]//2008 IEEE Conference on Computer Vision and Pattern Recognition. Anchorage, AK, USA, 2008: 1-8.
[5] GIRSHICK R, DONAHUE J, DARRELL T, et al.Rich feature hierarchies for accurate object detection and semantic segmentation[C]//2014 IEEE Conference on Computer Vision and Pattern Recognition. Columbus, OH, USA, 2014: 580-587.
[6] REN S Q, HE K M, 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.
[7] LIU W, ANGUELOV D, ERHAN D, et al.SSD: single shot MultiBox detector[C]//Computer Vision-ECCV 2016. Amsterdam, Netherlands, 2016: 21-37.
[8] REDMON J, DIVVALA S, GIRSHICK R, et al.You only look once: unified, real-time object detection[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas, NV, USA, 2016: 779-788.
[9] REDMON J, FARHADI A.YOLO9000: better, faster, stronger[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Honolulu, HI, USA, 2017: 6517-6525.
[10] BOCHKOVSKIY A, WANG C Y, LIAO H M. YOLOv4: optimal speed and accuracy of object detection[EB/OL].2020: arXiv: 2004.10934. https://arxiv.org/abs/2004.10934
[11] LI C Y, LI L L, JIANG H L, et al. YOLOv6: a single-stage object detection framework for industrial applications[EB/OL].2022: arXiv: 2209.02976. https://arxiv.org/abs/2209.02976
[12] 范钧玮, 饶全瑞, 赵薇, 等. 改进的YOLOv5双影像光伏故障小目标检测[J]. 太阳能学报, 2024, 45(7): 510-516.
FAN J W, RAO Q R, ZHAO W, et al.Improved YOLOv5 dual-image photovoltaic fault small target detection[J]. Acta energiae solaris sinica, 2024, 45(7): 510-516.
[13] 彭自然, 张颖清, 肖伸平. 基于YOLOv5的太阳电池表面缺陷检测[J]. 太阳能学报, 2024, 45(6): 368-375.
PENG Z R, ZHANG Y Q, XIAO S P.Research on surface defect detection of solar cell with improved YOLOv5 algorithm[J]. Acta energiae solaris sinica, 2024, 45(6): 368-375.
[14] ZHAO X L, SONG C H, ZHANG H F, et al.HRNet-based automatic identification of photovoltaic module defects using electroluminescence images[J]. Energy, 2023, 267: 126605.
[15] MENG Z Y, XU S Z, WANG L C, et al.Defect object detection algorithm for electroluminescence image defects of photovoltaic modules based on deep learning[J]. Energy science & engineering, 2022, 10(3): 800-813.
[16] 田浩, 周强, 贺晨龙. 基于多尺度特征融合的光伏组件缺陷检测[J]. 计算机工程与应用, 2024, 60(3): 340-347.
TIAN H, ZHOU Q, HE C L.Defect detection of photovoltaic modules based on multi-scale feature fusion[J]. Computer engineering and applications, 2024, 60(3): 340-347.
[17] CAO Y K, PANG D D, ZHAO Q C, et al.Improved YOLOv8-GD deep learning model for defect detection in electroluminescence images of solar photovoltaic modules[J]. Engineering applications of artificial intelligence, 2024, 131: 107866.
[18] LI Y X, HOU Q B, ZHENG Z H, et al.Large selective kernel network for remote sensing object detection[C]//2023 IEEE/CVF International Conference on Computer Vision (ICCV). Paris, France, 2024: 16748-16759.
[19] LI J F, WEN Y, HE L H.SCConv: spatial and channel reconstruction convolution for feature redundancy[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Vancouver, BC, Canada, 2023: 6153-6162.
[20] 白锋, 马庆禄, 赵敏. 面向航拍路面裂缝检测的AC-YOLO[J]. 计算机工程与应用, 2025, 61(1): 153-164.
BAI F, MA Q L, ZHAO M.AC-YOLO for aerial pavement crack detection[J]. Computer engineering and applications, 2025, 61(1): 153-164.
[21] ZHANG H, XU C, ZHANG S J. Inner-IoU: more effective intersection over union loss with auxiliary bounding box[EB/OL].2023: arXiv: 2311.02877. https://arxiv.org/abs/2311.02877.
[22] SU B Y, ZHOU Z, CHEN H Y.PVEL-AD: a large-scale open-world dataset for photovoltaic cell anomaly detection[J]. IEEE transactions on industrial informatics, 2023, 19(1): 404-413.
基金
光伏发电系统数字化评价技术研究（2022ZXJ04A03）; 黑龙江省自然科学基金项目（LH2023F003）
PDF(2802 KB)
