YOLO-DSBF：一种新的太阳电池缺陷识别方法

何羿颉; 楚瀛; 夏能弘; 江正源; 李曦

doi:doi:10.19912/j.0254-0096.tynxb.2024-1226

欢迎来到能源电力期刊网！

English Version 读者登录期刊登录

您当前的位置：

首页 >

文章列表页 >

YOLO-DSBF：一种新的太阳电池缺陷识别方法

更新时间：2025-12-10

- YOLO-DSBF：一种新的太阳电池缺陷识别方法
- YOLO-DSBF：一种新的太阳电池缺陷识别方法
- 太阳能学报 2025年第11期页码：280-288
- 作者机构：
- 作者简介：
- 基金信息：
- DOI：doi:10.19912/j.0254-0096.tynxb.2024-1226
  中图分类号： M914
- 网络出版：2025-12-04，
  
  纸质出版：2025-12-04
- 稿件说明：
移动端阅览
何羿颉, 楚瀛, 夏能弘, 等. YOLO-DSBF：一种新的太阳电池缺陷识别方法[J]. 太阳能学报, 2025,(11):280-288.

何羿颉, 楚瀛, 夏能弘, et al. YOLO-DSBF：一种新的太阳电池缺陷识别方法[J]. 2025, (11): 280-288.
何羿颉, 楚瀛, 夏能弘, 等. YOLO-DSBF：一种新的太阳电池缺陷识别方法[J]. 太阳能学报, 2025,(11):280-288. DOI： doi:10.19912/j.0254-0096.tynxb.2024-1226.

何羿颉, 楚瀛, 夏能弘, et al. YOLO-DSBF：一种新的太阳电池缺陷识别方法[J]. 2025, (11): 280-288. DOI： doi:10.19912/j.0254-0096.tynxb.2024-1226.

摘要

针对当前太阳电池缺陷检测模型存在识别精度低、检测不准确等问题

该文提出一种新的太阳电池缺陷识别方法YOLO-DSBF。首先

将YOLOv8头部网络中的C2f层融合动态蛇形卷积（DSConv）设计C2f_DSConv模块

使其能够自适应聚焦于细长局部缺陷;其次

在颈部网络中引入动态稀疏注意力机制（BiFormer）实现更灵活的内容识别和计算调配

提升模型的特征提取能力;然后

针对微小缺陷点

添加小目标检测层

降低漏检率;最后

采用GIoU损失函数替代原有的综合交并比（CIoU）损失函数

有效提升算法的回归性能。实验结果表明

相较于基准模型YOLOv8n

该模型mAP@0.5、mAP@0.5：0.95、精确度分别提升5.64%、5%、13.29%。该模型对比其他检测模型

在不增大模型尺寸的同时提高了检测精度

能更好的适用于太阳电池缺陷检测任务。

Abstract

Addressing the prevalent issues of low recognition accuracy and inaccurate detection in current solar cell defect detection models

this paper introduces a novel method

YOLO-DSBF

for solar cell defect detection. Firstly

the C2f layer in the YOLOv8 head network is integrated with Dynamic Snake Convolution （DSConv） to create the C2f_DSConv module

which enables adaptive focusing on elongated local defects. Secondly

a dynamic sparse attention mechanism （BiFormer） is incorporated into the neck network

facilitating more flexible computation allocation and content awareness

thereby enhancing the model's feature extraction capabilities. Subsequently

a small object detection layer is introduced to target micro-defects

effectively mitigating the missed detection rate. Lastly

replacing the original CIoU loss function with the GIoU loss function effectively enhances the regression performance of the algorithm. Experimental results reveal that

in comparison to the baseline model YOLOv8n

the proposed model exhibits improvements of 5.64%

and 13.29% in mAP@0.5

mAP@0.5：0.95

and precision

respectively. Without increasing the model size

this model surpasses other detection models in detection accuracy

rendering it more appropriate for solar cell defect detection tasks.

关键词

Keywords

references

DHIMISH M, HOLMES V.Solar cells micro crack detection technique using state-of-the-art electroluminescence imaging[J]. Journal of science: advanced materials and devices, 2019, 4(4): 499-508.

FUYUKI T, KITIYANAN A.Photographic diagnosis of crystalline silicon solar cells utilizing electroluminescence[J]. Applied physics A, 2009, 96(1): 189-196.

LI W C, TSAI D M.Wavelet-based defect detection in solar wafer images with inhomogeneous texture[J]. Pattern recognition, 2012, 45(2): 742-756.

黎亚娟, 黄伟, 周橤. 基于数学形态学的太阳能电池板划痕检测方法研究[J]. 机电技术, 2016, 39(1): 16-19.

JUMABOEV S, JURAKUZIEV D, LEE M.Photovoltaics plant fault detection using deep learning techniques[J]. Remote sensing, 2022, 14(15): 3728.

周颖, 毛立, 张燕, 等. 改进CNN的太阳电池缺陷识别方法研究[J]. 太阳能学报, 2020, 41(12): 69-76.

王道累, 李超, 李明山, 等. 基于深度卷积神经网络的光伏组件热斑检测[J]. 太阳能学报, 2022, 43(1): 412-417.

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.

CHEN H Y, PANG Y, HU Q D, et al.Solar cell surface defect inspection based on multispectral convolutional neural network[J]. Journal of intelligent manufacturing, 2020, 31(2): 453-468.

GE C P, LIU Z, FANG L M, et al.A hybrid fuzzy convolutional neural network based mechanism for photovoltaic cell defect detection with electroluminescence images[J]. IEEE transactions on parallel and distributed systems, 2021, 32(7): 1653-1664.

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.

孙海蓉, 刘永朋, 周黎辉. 基于轻量化YOLOv5s的光伏热斑检测定位方法[J]. 太阳能学报, 2024, 45(11): 282-288.

CHEN H Y, SONG M Y, ZHANG Z Z, et al.Detection of surface defects in solar cells by bidirectional-path feature pyramid group-wise attention detector[J]. IEEE transactions on instrumentation and measurement, 2022, 71: 5025609.

牛小育, 刘长良, 刘卫亮, 等. 基于深度学习的光伏红外图像热斑检测方法[J]. 太阳能学报, 2024, 45(11):272-281.

QI Y L, HE Y T, QI X M, et al.Dynamic snake convolution based on topological geometric constraints for tubular structure segmentation[C]//2023 IEEE/CVF International Conference on Computer Vision (ICCV). Paris, France, 2023: 6047-6056.

DAI J F, QI H Z, XIONG Y W, et al.Deformable convolutional networks[C]//2017 IEEE International Conference on Computer Vision (ICCV). Venice, Italy, 2017: 764-773.

HAN K, WANG Y H, TIAN Q, et al.GhostNet: more features from cheap operations[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle, WA, USA, 2020: 1577-1586.

ZHU L, WANG X J, KE Z H, et al.BiFormer: vision transformer with bi-level routing attention[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Vancouver, BC, Canada, 2023: 10323-10333.

DONG X Y, BAO J M, CHEN D D, et al.CSWin transformer: a general vision transformer backbone with cross-shaped windows[C]//2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). New Orleans, LA, USA, 2022: 12114-12124.

HUANG Z L, WANG X G, WEI Y C, et al.CCNet: criss-cross attention for semantic segmentation[J]. IEEE transactions on pattern analysis and machine intelligence, 2023, 45(6): 6896-6908.

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.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

基于HCF-YOLO的光伏组件红外图像热斑故障检测

基于多尺度渐近金字塔的太阳电池缺陷检测网络