针对传统红外光伏组件缺陷检测精确低、速度慢问题,提出一种基于改进YOLOv8的红外光伏组件缺陷检测方法WD_YOLOv8。首先引入Wise-IoU v2损失函数替换原算法中的CIoU损失,提升检测精确率,同时提高精确率P和召回率R之间的平衡;其次,在特征融合网络Neck中采用DySample上采样,保持较好的检测帧率。实验表明,改进后模型在红外伪彩色光伏组件缺陷检测场景下指标P、mAP@0.5和F1-score分别提高0.3、1.5和2.6个百分点;在红外灰度光伏组件缺陷检测场景下指标P、mAP@0.5和F1-score分别提高18.6、7.7和5.1个百分点。改进后算法在红外场景下具有良好的鲁棒性和适应性。
Abstract
Traditional infrared photovoltaic module defect detection methods often suffer from low accuracy and slow processing speeds. To address these limitations, we propose WD_YOLOv8, an improved infrared photovoltaic module defect detection method based on YOLOv8. First, we introduced the Wise Intersection over Union v2 loss function to replace the Complete-IoU loss function used in the original algorithm, thereby enhancing detection precision. Additionally, the v2 loss function improved the balance between precision （P） and recall （R）, optimizing overall detection performance. Second, DySample up-sampling was incorporated into the feature fusion network （Neck） to maintain a high detection frame rate. Experimental results show that the improved model increases metrics P, mAP@0.5, and F1-score by 0.3, 1.5, and 2.6 percentage points, respectively, in the infrared pseudo-color photovoltaic module defect detection scenarios. Moreover, in infrared grayscale photovoltaic module defect detection scenarios, these metrics improve by 18.6, 7.7, and 5.1 percentage points, respectively. These findings indicate that the proposed algorithm exhibits strong robustness and adaptability in infrared imaging applications.
关键词
光伏组件 /
红外热成像 /
表面缺陷 /
损伤检测 /
目标检测 /
转换效率 /
太阳电池
Key words
photovoltaic modules /
infrared thermal imaging /
surface defects /
damage detection /
object detection /
conversion efficiency /
solar cells
中图分类号：
TM615
TP183
TP391.41
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1] LIU W, ANGUELOV D, ERHAN D, et al.SSD: single shot MultiBox detector[C]//Computer Vision-ECCV 2016. Cham: Springer, 2016: 21-37.
[2] 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.
[3] HE K M, GKIOXARI G, DOLLÁR P, et al. Mask R-CNN[C]//2017 IEEE International Conference on Computer Vision (ICCV). Venice, Italy, 2017: 2980-2988.
[4] 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.
[5] 张上, 黄俊锋, 王恒涛, 等. 低空轻量级红外弱小目标检测算法[J]. 激光与红外, 2024, 54(1): 122-129.
ZHANG S, HUANG J F, WANG H T, et al.Low altitude lightweight infrared weak small target detection algorithm[J]. Laser & infrared, 2024, 54(1): 122-129.
[6] 沈凌云, 郎百和, 宋正勋, 等. 基于DCS-YOLOv8模型的红外图像目标检测方法[J]. 红外技术, 2024, 46(5): 565-575.
SHEN L Y, LANG B H, SONG Z X, et al.Infrared image object detection method based on DCS-YOLOv8 model[J]. Infrared technology, 2024, 46(5): 565-575.
[7] 李玉娟, 周界龙, 麦耀华, 等. YOLOv5用于太阳电池缺陷检测优化研究[J]. 太阳能学报, 2024, 45(11): 162-169.
LI Y J, ZHOU J L, MAI Y H, et al.YOLOv5 is used in optimization of surface defect detection of solar cells[J]. Acta energiae solaris sinica, 2024, 45(11): 162-169.
[8] 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.
[9] 徐威, 李为相, 方志, 等. 基于改进YOLOv7-tiny的光伏电池缺陷检测算法[J]. 计算机工程与应用, 2024, 60(15): 336-343.
XU W, LI W X, FANG Z, et al.Solar cell defect detection algorithm based on improved YOLOv7-tiny[J]. Computer engineering and applications, 2024, 60(15): 336-343.
[10] 廖力达, 罗晓, 黄斌, 等. 复杂背景下的多晶硅太阳电池缺陷检测[J]. 太阳能学报, 2024, 45(9): 295-303.
LIAO L D, LUO X, HUANG B, et al.Detection of defects in polysilicon solar cells in complex backgrounds[J]. Acta energiae solaris sinica, 2024, 45(9): 295-303.
[11] 陶志勇, 易廷军, 林森, 等. 轻量化M-CNN的太阳电池表面缺陷识别[J]. 太阳能学报, 2024, 45(6): 341-348.
TAO Z Y, YI T J, LIN S, et al.Lightweight M-CNN solar cell surface defectidentification[J]. Acta energiae solaris sinica, 2024, 45(6): 341-348.
[12] TONG Z J, CHEN Y H, XU Z W, et al. Wise-IoU: bounding box regression loss with dynamic focusing mechanism[EB/OL].2023: arXiv: 2301.10051. https://arxiv.org/abs/2301.10051
[13] LIU W Z, LU H, FU H T, et al.Learning to upsample by learning to sample[C]//2023 IEEE/CVF International Conference on Computer Vision (ICCV). Paris, France, 2024: 6004-6014.
[14] WANG J Q, CHEN K, XU R, et al.CARAFE: content-aware ReAssembly of FEatures[C]//2019 IEEE/CVF International Conference on Computer Vision (ICCV). Seoul, Korea, 2019: 3007-3016.
基金
安徽省高校自然科学研究重点项目（2022AH052740; 2023AH052297）; 安徽省职业与成人教育学会项目（AZCJ2024127）
PDF(2108 KB)
