FPC Surface Defect Detection Based on Improved YOLOv11n Model
GUAN Zhouyang
HUANG Yong
TAN Hao
TAN Hang
CHEN Jincai
Abstract：To address the challenges of high-precision recognition and high-speed processing required for the surface defect detection of flexible printed circuit board(FPC)in industrial production,an FPC surface defect detection model based on the improved you only look once version 11 nano(YOLOv11n)was proposed.The main improvements were as follows:a slim-neck module was introduced to replace the convolution(Conv)layers in the original YOLOv11n neck network with ghost shuffle convolution(GSConv),and the convolutional three-scale kernel-adaptive dual-path(C3K2)structure was replaced with vortex of vectorized ghost shuffle cross stage partial(VoVGSCSP)structure;auxiliary detection heads were simultaneously introduced to enhance feature extraction capabilities,and focaler-complete intersection over union(Focaler-CIoU)was adopted to improve the original loss function.The results showed that compared with the original YOLOv11n model,the improved YOLOv11n model's recall rate and mean average precision were increased by 2.9 and 2.6 percent respectively,the detection speed reached 116.5 frames/s,and the model parameters were reduced by approximately 0.49%.The model was able to effectively achieve high-precision detection of small-sized and dense defects on FPC surfaces,meeting the dual requirements of precision and real-time performance in industrial production,and providing a reliable solution for quality control in FPC manufacturing.
Keywords：flexible printed circuit boardsurface defect detectionYOLOv11nslim-neckauxiliary detection headloss function
Publication Date：2025-06-20
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：8( 210-216,300 )