摘要
本文提出了一种基于改进YOLOv8的电路板缺陷检测方法,旨在提高电路板缺陷的检测精度和泛化能力。通过对YOLOv8算法进行优化,包括引入MPDIoU损失函数优化定位精度、在主干网络中引入C2fBF模块增强特征提取能力、在Neck端添加浅层特征图并使用GSConv和VoVGSCSPC模块进行特征融合,以及采用多种数据增强技术和自动化超参数优化工具,本文方法显著提升了检测性能。此外,本文还采用了经典的非极大值抑制算法和基于梯度下降的优化算法进行后处理,进一步提高了检测结果的准确性和稳定性。实验结果表明,该方法在电路板缺陷检测任务中具有优异的表现,为电路板质量检测和自动化生产提供了有力的技术支持。
This article proposes a circuit board defect detection method based on improved YOLOv8,aiming to improve the detection accuracy and generalization ability of circuit board defects.By optimizing the YOLOv8 algorithm,including introducing MPDIoU loss function to improve localization accuracy,introducing C2fBF module in the backbone network to enhance feature extraction capability,adding shallow feature maps on the Neck end and using GSConv and VoVGSCSPC modules for feature fusion,and adopting various data augmentation techniques and automated hyperparameter optimization tools,this method significantly improves detection performance.In addition,this article also uses classic non maximum suppression algorithms and gradient descent based optimization algorithms for post-processing,further improving the accuracy and stability of the detection results.The experimental results show that this method has excellent performance in circuit board defect detection tasks,providing strong technical support for circuit board quality inspection and automated production.
作者
林荣昊
易子豐
贺梓修
程嵩岐
LIN Rong-hao;YI Zi-feng;HE Zi-xiu;CHENG Song-qi(Guangzhou College of Applied Science and Technology,Zhaoqing 511370,China;Guangdong Xizhen Circuit Technology Co.,Ltd.,Zhaoqing 511370,China)
出处
《价值工程》
2025年第26期24-26,共3页
Value Engineering
