基于YOLO11的轻量化PCB缺陷检测算法研究

doi:10.16088/j.issn.1001-6600.2025022502

摘要/Abstract

摘要： 针对印刷电路板(printed circuit board, PCB)小目标缺陷检测精度低,且模型复杂、计算量大、难以在边缘设备上部署运行的问题,本文基于YOLO11n提出一种轻量化算法。首先使用BiMAFPN(bi-directional multi-branch auxiliary feature pyramid network)对模型的网络结构进行重构,再使用C3k2_Faster模块在保证准确度的前提下进一步降低模型的复杂度,最后使用LSCD(lightweight shared convolutional detection)检测头提高检测的精度。实验表明,本文提出的模型精确率达93.0%,召回率82.8%,模型权重大小3.8 MiB,mAP@0.5和mAP@0.5∶0.95分别达到89.9%和47.1%,相较于YOLO11n,精确率提升0.6个百分点,mAP@0.5和mAP@0.5∶0.95分别提升1.4和0.6个百分点,模型体积、计算量和参数量分别减少30.9%、19.0%、34.6%。改进后的算法在轻量化的同时仍具备较好的检测精度,适合于边缘设备的部署。

关键词: YOLO11, PCB缺陷, 轻量化, BiFPN, 目标检测

Abstract: To address the issues of low detection accuracy, high model complexity, and excessive computational costs in small-target defect detection of printed circuit boards (PCBs), which hinder deployment on edge devices, a lightweight algorithm based on YOLO11n was proposed. Firstly, the BiMAFPN (Bi-Directional Multi-Branch Auxiliary Feature Pyramid Network) architecture is employed to reconstruct the network structure. Subsequently, the C3k2_Faster module is implemented to reduce model complexity while maintaining detection accuracy. Finally, the LSCD (Lightweight Shared Convolutional Detection) head is introduced to enhance precision. Experimental results demonstrate that the proposed model achieves 93.0% precision and 82.8% recall, with a compact model size of 3.8 MiB. Enhancements include a 0.6 percentage points increase in precision. The mean average precision (mAP) values reach 89.9% (mAP@0.5) and 47.1% (mAP@0.5:0.95), representing improvements of 1.4 and 0.6 percentage points respectively compared with the baseline YOLO11n model while reducing model size, computational complexity, and parameter count by 30.9%, 19.0% and 34.6% respectively. These optimizations enable the improved algorithm to maintain competitive detection performance while achieving significant lightweight characteristics, demonstrating strong potential for practical deployment in edge computing environments.

Key words: YOLO11, PCB defect, lightweight, BiFPN, object detection

中图分类号: TP391.41

黄文杰, 罗维平, 陈镇南, 彭志祥, 丁梓豪. 基于YOLO11的轻量化PCB缺陷检测算法研究[J]. 广西师范大学学报（自然科学版）, 2026, 44(1): 56-67.

HUANG Wenjie, LUO Weiping, CHEN Zhennan, PENG Zhixiang, DING Zihao. Research on Lightweight PCB Defect Detection Algorithm Based on YOLO11[J]. Journal of Guangxi Normal University(Natural Science Edition), 2026, 44(1): 56-67.

参考文献

[1] KUO C F J, FANG T Y, LEE C L, et al. Automated optical inspection system for surface mount device light emitting diodes[J]. Journal of Intelligent Manufacturing, 2019, 30(2): 641-655. DOI: 10.1007/s10845-016-1270-6.
[2] LOU L S, LU K, XUE J. Defect detection based on improved YOLOx for ultrasonic images[J]. Sensing and Imaging, 2024, 25(1): 10. DOI: 10.1007/s11220-024-00459-4.
[3] 郭渊, 许伟佳, 董振标. 基于深度学习的PCB表面缺陷检测研究进展[J]. 机床与液压, 2024, 52(23): 188-198. DOI: 10.3969/j.issn.1001-3881.2024.23.030.
[4] 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). Los Alamitos, CA: IEEE Computer Society, 2016: 779-788. DOI: 10.1109/CVPR.2016.91.
[5] LIU W, ANGUELOV D, ERHAN D, et al. SSD: single shot multibox detector[C]//Computer Vision-ECCV 2016: LNCS Volume 9905.Cham: Springer International Publishing AG, 2016: 21-37. DOI: 10.1007/978-3-319-46448-0.
[6] TAN M X, PANG R M, LE Q V. EfficientDet: scalable and efficient object detection[C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2020: 10778-10787. DOI: 10.1109/CVPR42600.2020.01079.
[7] 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. DOI: 10.1109/TPAMI.2016.2577031.
[8] 李俊. 基于SSD模型的PCB外观缺陷检测[J]. 现代计算机, 2022, 28(7): 79-82, 91. DOI: 10.3969/j.issn.1007-1423.2022.07.014.
[9] 吴伟浩, 李青. 基于改进Yolo v3的电连接器缺陷检测[J]. 传感技术学报, 2020, 33(2): 299-307. DOI: 10.3969/j.issn.1004-1699.2020.02.022.
[10] 龚陈博, 南卓江, 陶卫. 基于改进Faster RCNN的PCB表面缺陷检测研究[J]. 自动化仪表, 2024, 45(7): 99-103, 109. DOI: 10.16086/j.cnki.issn1000-0380.2023060097.
[11] 姜源, 付波, 权轶, 等. 基于改进YOLOv8n的PCB缺陷检测算法[J]. 国外电子测量技术, 2024, 43(6): 22-32. DOI: 10.19652/j.cnki.femt.2305887.
[12] 王健, 肖迪, 冯李航, 等. 基于改进YOLOv8s的PCB小目标缺陷检测模型[J]. 计算机工程与应用, 2025, 61(15): 288-297. DOI: 10.3778/j.issn.1002-8331.2406-0020.
[13] 吕秀丽, 杨昕升, 曹志民. 改进YOLOv8的PCB表面缺陷检测算法[J]. 电子测量技术, 2024, 47(12): 100-108. DOI: 10.19651/j.cnki.emt.2415797.
[14] 张淑卿, 孟昊, 葛超. 改进YOLOv8n的轻量级PCB缺陷检测算法[J]. 现代电子技术, 2024, 47(15): 115-121. DOI: 10.16652/j.issn.1004-373x.2024.15.019.
[15] CHAN K Y, YIU K F C, LAM H K, et al. Ball bonding inspections using a conjoint framework with machine learning and human judgement[J]. Applied Soft Computing, 2021, 102: 107115. DOI: 10.1016/j.asoc.2021.107115.
[16] 李忠科, 刘小芳. 基于轻量级YOLOv8n网络的PCB缺陷检测算法[J]. 电子测量技术, 2024, 47(4): 120-126. DOI: 10.19651/j.cnki.emt.2315076.
[17] 王军, 伍毅, 陈正超. 基于SMT-YOLOv8的PCB缺陷检测研究[J]. 电子测量技术, 2024, 47(11): 131-137. DOI: 10.19651/j.cnki.emt.2416003.
[18] YANG Z Q, GUAN Q, ZHAO K E, et al. Multi-branch auxiliary fusion YOLO with re-parameterization heterogeneous convolutional for accurate object detection[C]//Pattern Recognition and Computer Vision: LNCS Volume 15042. Singapore: Springer Nature Singapore Pte Ltd., 2024: 492-505. DOI: 10.1007/978-981-97-8858-3_34.
[19] CHEN J R, KAO S H, HE H, et al. Run, don’t walk: chasing higher FLOPS for faster neural networks[C]//2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Los Alamitos, CA: IEEE Computer Society, 2023: 12021-12031. DOI: 10.1109/CVPR52729.2023.01157.
[20] 周翔, 王可庆, 周新翔, 等. 基于改进YOLOv10n的电动车头盔佩戴检测算法[J]. 电子测量技术, 2025, 48(5): 40-49. DOI: 10.19651/j.cnki.emt.2417353.
[21] 王海群, 王炳楠, 葛超. 改进ESP-YOLO的PCB缺陷检测算法[J]. 计算机工程与科学, 2025, 47(2): 317-326. DOI: 10.3969/j.issn.1007-130X.2025.02.014.
[22] 颜豪男, 吕伏. 多尺度特征增强的PCB板表面缺陷检测算法[J/OL]. 信息与控制: 1-13[2025-02-25]. https://doi.org/10.13976/j.cnki.xk.2024.4021.
[23] 范博淦, 王淑青, 陈开元. 基于改进YOLOv8的印刷电路板缺陷检测模型[J]. 现代电子技术, 2025, 48(11): 144-150. DOI: 10.16652/j.issn.1004-373x.2025.11.022.
[24] LIU C J, ZHANG M X, YAN H W, et al. CSYOLO: a YOLOv8-based PCB defect detection model integrating composite backbone networks and dynamic snake convolution[J/OL]. Journal of Measurement Science and Instrumentation: 1-12[2025-06-28]. https://link.cnki.net/urlid/14.1357.th.20250324.1743.002.

Metrics

Viewed

Full text

Abstract

Cited

Shared

Discussed