电子测量与仪器学报

钱万明,朱红萍,朱泓知,李毅,郭利文.基于自适应加权特征融合的 PCB 裸板缺陷检测研究[J].电子测量与仪器学报,2022,36(10):92-99

基于自适应加权特征融合的 PCB 裸板缺陷检测研究

Research on defect detection of PCB bare board basedon adaptive weighted feature fusion

DOI：

英文关键词:defect detection lightweight network bidirectional adaptive feature fusion m-ECANet

基金项目:国家自然科学基金（61973109）项目资助

作者	单位
钱万明	1.湖南科技大学信息与电气工程学院
朱红萍	1.湖南科技大学信息与电气工程学院
朱泓知	1.湖南科技大学信息与电气工程学院
李毅	1.湖南科技大学信息与电气工程学院
郭利文	1.湖南科技大学信息与电气工程学院

Author	Institution
Qian Wanming	1.School of Electrical Information Engineering, Hunan University of Science and Technology
Zhu Hongping	1.School of Electrical Information Engineering, Hunan University of Science and Technology
Zhu Hongzhi	1.School of Electrical Information Engineering, Hunan University of Science and Technology
Li Yi	1.School of Electrical Information Engineering, Hunan University of Science and Technology
Guo Liwen	1.School of Electrical Information Engineering, Hunan University of Science and Technology

摘要点击次数: 520

全文下载次数: 656

中文摘要:

现有的针对 PCB 裸板的缺陷检测方法存在精确度低、实时性差且难以在移动端部署等问题,本文以 YOLO( you only look once)v4 算法为基本框架并对其进行改进,提出了一种专门针对 PCB 裸板的缺陷检测算法。针对 YOLOv4 算法难以在移动端部署的问题,采用 GhostNet 取代 CSPDarknet53 以轻量化整个检测网络。为弥补 YOLOv4 算法在多尺度特征融合方面的性能不足,提出了一种双向自适应特征融合网络 AF-BiFPN 取代 PANet 网络。为进一步提高模型的检测精度,在 AF-BiFPN 特征融合网络的采样的过程中插入 m-ECANet 通道注意力机制。实验结果证明,改进后的 YOLOv4 算法的模型大小为 18. 64 MB,检测的平均精度(mean average precision,mAP)为 98. 39%,检测速度为 62. 23 FPS,可为实际 PCB 裸板检测提供理论指导。

英文摘要:

The existing defect detection methods for PCB bare boards have problems such as low accuracy, poor real-time performance, and difficulty in deploying on mobile terminals. Based on the YOLOv4 algorithm as the basic framework to improve it, this paper proposes a defect detection algorithm specifically for PCB bare boards. In response to the problem that the YOLOv4 algorithm is difficult to deploy on the mobile terminal, the proposed improved algorithm uses GhostNet instead of CSPDarknet53 to lighten the entire detection network. In order to make up for the lack of performance of YOLOv4 algorithm in multi-scale feature fusion, this paper proposes a bidirectional adaptive feature fusion network AF-BiFPN to replace the PANet network in YOLOv4 algorithm. In order to further improve the detection accuracy of the model, the m-ECANet channel attention mechanism is inserted in the sampling process of the AF-BiFPN feature fusion network. The experimental results show that the model size of the improved YOLOv4 algorithm is 18. 64 MB, the mean average precision (mAP) of detection is 98. 39%, and the detection speed is 62. 23 FPS, which can provide theoretical guidance for actual PCB bare board detection.

查看全文查看/发表评论下载PDF阅读器