[1]刘 骞,陈茂林.基于深度学习的少样本光伏边框划痕检测[J].机械与电子,2025,(08):47-53.
　LIU Qian,CHEN Maolin.Few-shot Scratch Detection on Photovoltaic Frame Based on Deep Learning[J].Machinery & Electronics,2025,(08):47-53.
点击复制

基于深度学习的少样本光伏边框划痕检测()
分享到：

《机械与电子》[ISSN:1001-2257/CN:52-1052/TH]

卷:
期数:
2025年08期
页码:
47-53
栏目:
自动控制与检测
出版日期:
2025-08-25

文章信息/Info

Title:
Few-shot Scratch Detection on Photovoltaic Frame Based on Deep Learning
文章编号:
1001-2257 ( 2025 ) 08-0047-07
作者:
刘 骞; 陈茂林
同济大学机械与能源工程学院,上海 201804
Author(s):
LIU Qian ;  CHEN Maolin
( The School of Mechanical Engineering , Tongji University , Shanghai 201804 , China )
关键词:
缺陷检测; 深度学习; k-means 聚类;  SPPFCSPC ;  AKConv ; 损失函数
Keywords:
defect detection ;  deep learning ;  k means clustering ;  SPPFCSPC ;  AKConv ;  loss function
分类号:
TP391.4
文献标志码:
A
摘要:
针对光伏板铝合金边框表面划痕检测中存在的小样本、背景复杂等问题,提出了一种基于改进 YOLOv5s 的深度学习检测方法。先通过 k means 聚类算法训练锚框数据,再引入 SPPFCSPC 模块,融合 AKConv 卷积,并采用 Shape IoU 损失函数与 Soft NMS 算法。实验选用 73 张工业现场采集的划痕图像(训练集 66 张,验证集 7 张),在有限算力环境下进行训练。结果表明,改进后的 YOLOv5s KSASS 模型在平均精度、精确率和召回率上分别达到 0.932 11 、 0.999 75 和 0.857 14 ,较原始 YOLOv5s 模型提升了126.3% 、16.2% 和 100.7% ,有效解决了小样本条件下复杂背景干扰和微弱缺陷检测难题,为工业场景中的高精度表面缺陷检测提供了轻量化解决方案。未来将进一步优化模型对低对比度划痕的敏感性,并扩展至多类别缺陷检测任务。
Abstract:
To address the issues of small sample size and complex backgrounds in scratch detection on photovoltaic panel aluminum frames , this paper proposes a detection method based on an improved YOLOv5s.Firstly , the anchor box data is trained through the k-means clustering algorithm.Then , the SPPFCSPC module is introduced , the AKConv convolution is integrated , and the Shape-IoU loss function and the Soft-NMS algorithm are adopted.Experiments were conducted with 73 scratch images collected in an industrial setting ( 66 for training , 7 for validation ) in a limited computing environment.Results show that the improved YOLOv5s KSASS model achieves 0.932 11 mAP@0.5 , 0.999 75 precision , and 0.857 14 recall , representing increases of 126.3% , 16.2% , and 100.7% over the original YOLOv5s.This effectively solves the problem of detecting weak defects under complex backgrounds with small samples , offering a lightweight solution for high precision industrial surface defect detection.Future work will focus on enhancing sensitivity to low-contrast scratches and expanding to multi-category defect detection.

参考文献/References:

[ 1 ] 张利平,张红英,吴斌 . 基于多种边缘检测的视频划痕检测技术[ J ] . 电视技术, 2010 , 34 ( 1 ): 85-87.

[ 2 ] 张益翔 . 基于小波变换和 LBP 的木材表面缺陷识别[ D ] . 南京:南京林业大学,2017.
[ 3 ] 陈宗仁,谢文达,余君,等 . 基于深度学习的金属机械零件表面缺陷检测方法 [ J ] . 制造业自动化, 2021 , 43( 12 ): 170-173.
[ 4 ] QIU K P , TIAN L , WANG P. An effective framework of automated visual surface defect detection for metal parts [ J ] .IEEE Sensors journal , 2021 , 21 ( 8 ): 20412-20420.
[ 5 ] LIU Z X , HU Z L , WANG L X , et al.Effective detection of metal surface defects based on double-line laser ultrasonic with convolutional neural networks [ J ] .Modern physics letters B , 2021 , 35 ( 15 ): 2150263.
[ 6 ] 张海刚,鲁伽祎,匡国文,等 . 基于孪生网络的工业缺陷弱监督视觉检测算法 [ J ] . 深圳大学学报 (理工版),2023 , 40 ( 6 ): 657-664.
[ 7 ] HE D , XU K , ZHOU P. Defect detection of hot rolled steels with a new object detection framework called classification priority network [ J ] .Computers and industrial engineering , 2019 , 128 : 290-297.
[ 8 ] JIANG Z H , DOU X H , LIU X L , et al.Internal thread defect generation algorithm and detection system based on generative adversarial networks and you only look once [ J ] .Sensors , 2024 , 24 ( 17 ): 5636.
[ 9 ] 王一,龚肖杰,程佳,等 . 基于改进 YOLOv5 的金属工件表面缺陷检测[ J ] . 包装工程, 2022 , 43 ( 15 ): 54-60.
[ 10 ] 余松森,张明威,杨欢 . 改进 YOLOv5 的瓷砖表面缺陷检测[ J ] . 计算机系统应用, 2023 , 32 ( 8 ): 151-161.
[ 11 ] 栗庆杰 . 启发式 k-means 聚类算法的改进与应用研究[ D ] . 大连:大连交通大学,2023.
[ 12 ] XIE Y , YIN B , HAN X , et al.Improved YOLOv7 based steel surface defect detection algorithm [ J ] . Mathematical biosciences and engineering , 2024 , 21( 1 ): 346-368.
[ 13 ] SUN H , TAN C H , PANG S , et al.RA-YOLOv8 : an improved YOLOv8 seal text detection method [ J ] . Electronics , 2024 , 13 ( 15 ): 3001.
[ 14 ] ZHAO L , FU L L , JIA X , et al.YOLO-BOS : an emerging approach for vehicle detection with a novel BRSA mechanism [ J ] .Sensors , 2024 , 24 ( 24 ): 8126.

相似文献/References:

[1]王骁贤,张保华,陆思良.基于连续小波变换和卷积神经网络的无刷直流电机故障诊断[J].机械与电子,2018,(06):29.
　WANG Xiaoxian,ZHANG Baohua,LU Siliang.Fault Diagnosis of Brushless Direct Current Motor Based on Continuous Wavelet Transform and Convolutional Neural Network[J].Machinery & Electronics,2018,(08):29.
[2]刘志宇,黄亦翔.基于深度学习和迁移学习的液压泵健康评估方法[J].机械与电子,2018,(09):67.
　LIU Zhiyu,HUANG Yixiang.Health Assessment for Hydraulic Pump Based on Deep Learning and Transfer Learning[J].Machinery & Electronics,2018,(08):67.
[3]肖倩宏,康 鹏,杜 江,等.深度学习在电网智能调控系统中应用研究[J].机械与电子,2021,(01):38.
　XIAO Qianhong,KANG Peng,DU Jiang,et al.Research on the Application of Deep Learning Theory in Power Grid Intelligent Dispatching[J].Machinery & Electronics,2021,(08):38.
[4]许 哲,张少帅,郭 璐,等.无人机深度学习去雾算法[J].机械与电子,2021,(04):13.
　XU Zhe,ZHANG Shaoshuai,GUO Lu,et al. Deep Learning Defogging Algorithm for UAV[J].Machinery & Electronics,2021,(08):13.
[5]齐爱玲１,李 琳１,朱亦轩２,等.基于融合特征的双通道CNN滚动轴承故障识别[J].机械与电子,2021,(05):15.
　QI Ailing,LI Lin,ZHU Yixuan,et al.Dual Channel CNN Bearing Fault Identification Based on Fusion Feature[J].Machinery & Electronics,2021,(08):15.
[6]徐先峰,郑少杰,赵 依,等.基于数据分解与重构的光伏发电功率超短期预测[J].机械与电子,2022,(04):20.
　XU Xianfeng,ZHENG Shaojie,ZHAO Yi,et al.Ultra-short-term Prediction of Photovoltaic Power Generation Based on Data Decomposition and Deconstruction[J].Machinery & Electronics,2022,(08):20.
[7]江 励,熊达明,汤健华,等.自然光线环境中的空间物体快速识别和定位算法研究[J].机械与电子,2022,(06):8.
　JIANG Li,XIONG Daming,TANG Jianhua,et al.Recognition and Positioning Algorithm of Space Objects in Natural Light Environment[J].Machinery & Electronics,2022,(08):8.
[8]王西志,管声启,张理博,等.基于视觉引导的工业棒材上料系统研究[J].机械与电子,2023,41(05):19.
　WANG Xizhi,GUAN Shengqi,ZHANG Libo,et al.Research on Industrial Bar Feeding System Based on Visual Guidance[J].Machinery & Electronics,2023,41(08):19.
[9]王 青,吕绪山,党 帅,等.基于深度学习的纱管识别方法研究[J].机械与电子,2023,41(12):20.
　WANG Qing,LYU Xushan,DANG Shuai,et al.Research on Yarn Bobbin Detection Method Based on Deep Learning[J].Machinery & Electronics,2023,41(08):20.
[10]陆 晓,高 超,蒋承伶,等.基于双注意力机制和级联检测框架的金具螺栓缺陷检测研究[J].机械与电子,2024,42(01):63.
　LU Xiao,GAO Chao,JIANG Chengling,et al.Research on Bolt Defect Detection Based on Dual-attention Mechanism and Cascade Detection Framework[J].Machinery & Electronics,2024,42(08):63.

备注/Memo

备注/Memo:
收稿日期: 2025-03-18
作者简介:刘 骞 ( 2002- ),男,湖南邵阳人,硕士研究生,研究方向为机械设计理论、缺陷检测;陈茂林 ( 1977- ),男,湖北黄冈人,博士,讲师,硕士研究生导师,研究方向为机械设计理论、工程机械、智能设备和轨道车辆制动系统,通信作者, E-mail : forestchen@tongji.edu.cn 。
更新日期/Last Update: 2025-09-05