三峡大学水电工程智能视觉监测湖北省重点实验室;三峡大学湖北省建筑质量检测装备工程技术研究中心;三峡大学计算机与信息学院;荆楚理工学院大数据研究中心;
针对印刷电路板(Printed Circuit Board, PCB)表面缺陷检测任务中模型体积和参数量较大的问题，提出了一种基于通道剪枝的轻量级YOLOv8n网络PCB缺陷检测算法。为有效提升对PCB小目标缺陷的特征提取能力，采用RepViT作为特征提取网络；为提升网络对小目标的关注度，减少神经网络推理过程中的梯度信息重复，将颈部网络的卷积模块替换为Rep-Net with Cross-Stage Partial CSP and ELAN(RepNCSPELAN4);为降低缺陷重叠时检测框失真现象，在预测部分使用Focaler-MPDIoU替换完全交并比(Complete Intersection over Union, CIoU);利用层自适应幅度分数剪枝(Layer Adaptive Magnitude based Pruning, LAMP)方法对融合改进方法的模型进行修剪，去除模型中冗余的梯度信息和权重，减少参数量和浮点运算量，压缩模型体积。实验结果表明，在PCB公开数据集中，经过LAMP之后，该算法相较于YOLOv8n,参数量下降60.8%,模型体积减小50.8%,计算量下降48.8%,平均精度均值(mean Average Precision, mAP)提高3.8%。在提高精度的同时，计算量、参数量和模型体积都低于原模型，满足在低配置设备下的使用需求。
| 1,125 | 0 | 59 |
| 下载次数 | 被引频次 | 阅读次数 |
[1] 陆维宽，周志立，阮秀凯，等.基于注意力机制与多尺度融合的PCB缺陷检测[J].无线电工程，2024,54(1):6-13.
[2] LIU W,ANGUELOV D,ERHAN D,et al.SSD:Single Shot Multibox Detector[C]//Computer Vision-ECCV2016:14th European Conference.Amsterdam:Springer,2016:21-37.
[3] JIANG P Y,ERGU D,LIU F Y,et al.A Review of Yolo Algorithm Developments[J].Procedia Computer Science,2022,199:1066-1073.
[4] HE K M,GKIOXARI G,DOLLáR P,et al.Mask R-CNN[C]//Proceedings of the IEEE International Conference on Computer Vision.Venice:IEEE,2017:2961-2969.
[5] GIRSHICK R.Fast R-CNN[C]//Proceedings of the IEEE International Conference on Computer Vision.Santiago:IEEE,2015:1440-1448.
[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,2016,39(6):1137-1149.
[7] XIN H J,CHEN Z B,WANG B Y.PCB Electronic Component Defect Detection Method Based on Improved YOLOv4 Algorithm[C]// 6th International Conference on Electronic Technology and Information Science (ICETIS 2021).Harbin:IOP Publishing Ltd.,2021:012167.
[8] LI J,GU J N,HUANG Z D,et al.Application Research of Improved YOLO V3 Algorithm in PCB Electronic Component Detection[J].Applied Sciences,2019,9(18):3750.
[9] HU B,WANG J H.Detection of PCB Surface Defects with Improved Faster-RCNN and Feature Pyramid Network[J].IEEE Access,2020,8:108335-108345.
[10] LIN Q Q,ZHOU J Z,MA Q R,et al.EMRA-Net:A Pixel-wise Network Fusing Local and Global Features for Tiny and Low-contrast Surface Defect Detection[J].IEEE Transactions on Instrumentation and Measurement,2022,71:2504314.
[11] 朱瑞，齐元胜，张勇斌.融合部分卷积和无参数注意力机制的轻量级印刷电路板缺陷检测算法[J/OL].计算机集成制造系统，1-18[2024-04-24].https://doi.org/10.13196/j.cims.2023.0407.
[12] 彭辉，周博文，欧阳婉卿，等.基于双通道注意力的轻量化PCB缺陷检测研究[J].光电子·激光，2024,35(5):506-515.
[13] SHE P,LI X,YANG N,et al.Lightweight YOLOv8 PCB Defect Detection Algorithm Based on Triple Attention[J].Microelectronics & Computer,2024,41(4):20-30.
[14] WANG A,CHEN H,LIN Z J,et al.Rep ViT:Revisiting Mobile CNN from ViT Perspective[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition.Seattle:IEEE,2024:15909-15920.
[15] HOWARD A,SANDLER M,CHEN B,et al.Searching for MobileNetV3[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision.Seoul:IEEE,2019:1314-1324.
[16] 张果，陈逃，王剑平，等.轻量化的PCB表面缺陷检测算法[J].北京邮电大学学报，2024,47(2):38-44.
[17] WANG C Y,YEH I H,LIAO H Y M.YOLOv9:Learning What You Want to Learn Using Programmable Gradient Information[EB/OL].(2024-02-21) [2024-04-24].https://arxiv.org/abs/2402.13616.
[18] LEE J H,PARK S J,MO S W,et al.Layer-adaptive Sparsity for the Magnitude-based Pruning[EB/OL].(2020-10-15) [2024-05-04].https://arxiv.org/abs/2010.07611.
[19] ZHANG H,ZHANG S J.Focaler-IoU:More Focused Intersection over Union Loss[EB/OL].(2024-01-19) [2024-05-11].https://arxiv.org/abs/2401.10525.
[20] MA S L,XU Y.MPDIoU:A Loss for Efficient and Accurate Bounding Box Regression[EB/OL].(2023-07-14) [2024-06-02].https://arxiv.org/abs/2307.07662.
[21] HUANG W B,WEI P.A PCB Dataset for Defects Detection and Classification[EB/OL].(2019-06-24) [2024-06-14].https://arxiv.org/abs/1901.08204.
基本信息:
DOI：
中图分类号:TN41;TP391.41
引用信息:
[1]杨慧聪,陈慈发,张上.基于通道剪枝的YOLOv8n印刷电路板缺陷检测[J].无线电工程,2025,55(02):271-280.
基金信息:
湖北省大学生创新创业训练计划(202311075046); 国家级大学生创新创业训练计划(202111075012,202011075013)~~