Research on PCB defect detection algorithm based on mixed attention mechanism
DOI:
https://doi.org/10.54097/hset.v7i.1026Keywords:
Deep learning, Attention mechanism, ECANetAbstract
The production process of PCB is complex and cumbersome, which leads to the complex and diverse defects of the PCB and seriously affects the productivity. Therefore, timely and effective detection of its surface defects is particularly momentous. For this article, a new mixed attention module C_Efficient Channel Attention (C_ECA) is presented by connecting spatial attention with effective Channel attention. A one-dimensional convolution can be substituted for the fully connected layer in the channel attention module of C_ECA; for the spatial attention module of C_ECA, we can use 3*3 and 7*7 convolution kernel in place of the 7*7 convolution kernel to use multiple scale information to consider the significance of diverse regional trait. We integrated the module into YOLOv3 and tested it on PCB data set. Finding in the laboratory indicate that the network performance can be greatly improved and PCB defects can be accurately identified.
Downloads
References
Jaderberg M, Simonyan K, Zisserman A. Spatial transformer networks [C] //Advances in neural information processing systems. 2015: 2017-2025. J. Clerk Maxwell, A Treatise on Electricity and Magnetism, 3rd ed., vol. 2. Oxford: Clarendon, 1892, pp.68–73.
Fu J, Zheng H, Mei T. Look Closer to See Better: Recurrent Attention Convolutional Neural Network for Fine-Grained Image Recognition[C]// IEEE Conference on Computer Vision &Pattern Recognition. IEEE, 2017: 4438- 4446.
Girshick R, Donahue J, Darrell T, et al. Rich feature hierarchies for accurate object detection and semantic 7 segmentation[C]//Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, 2014, 580-587.
Girshick R. Fast r-cnn[C]//Proceedings of the IEEE international conference on computer vision. 2015: 1440-1448.
Ren S, He K, Girshick R, et al. Faster r-cnn: Towards real-time object detection with region proposal networks[C]//Advances in neural information processing systems. 2015: 91-99.
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.
Tsai D M,Chou Y H. Fast and precise positioning in PCBs using deep neural network regression[J].IEEE Transactions on Instrum entation and Measurement,2020,69(7): 4692-4701.
Liu W, Anguelov D, Erhan D, et al. SSD: Single shot multibox detector [C] //European conference on computer vision. Springer, Cham, 2016: 21-37.
Redmon J, Divvala S, Girshick R, et al. You Only look once: unified, real time object detection[C]//Computer Vision and Pattern Recognition. 2017: 6517-6525.
Redmon J, Farhadi A. YOLO9000: Better, faster, stronger [C] //IEEE conference on Computer Vision and Pattern Recognition, 2017: 6517-6525.
Redmon J, Farhadi A. YOLOv3: An incremental improvement [C] //IEEE conference on Computer Vision and Pattern Recognition, 2018, arXiv: 1804.0276.
Bochkovskiy A, Wang C Y, Liao H. YOLOv4: Optimal Speed and Accuracy of Object Detection[J]. 2020.
Li Y T, Kuo P, Guo J N. Automatic industry PCB board DIP process defect detection with deep ensemble method[C]// 2020 IEEE 29th International Symposium on Industrial Electronics (ISIE). Delft, Netherlands: IEEE, 2020: 453-459.
WOO S, PARK J, LEE J Y, et al. CBAM, convolutional block attention module[J]. arXiv, 1807.06521, 2018.
HU J, SHEN L, SUN G, et al. Squeeze- and- excitation networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2020, 42(8), 2011-2023.
PARK J, WOO S, LEE J Y, et al. A simple and lightweight attention module for convolutional neural networks[J]. International Journal of Computer Vision, 2020, 128(9), 783-798.
LI X, WANG W, HU X, et al. Selective kernel networks[C]// IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2020.
WANG Q, WU B, ZHU P, et al. ECA-Net, efficient channel attention for deep convolutional neural networks[C]// IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2020.
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
