摘要
随着电子信息制造业的快速发展,电路板质量要求日益提高,传统的人工检测方法难以满足高精度、高效率的检测需求。因此,本文提出一种基于卷积神经网络(CNN)的PCB板图像缺陷检测方法。首先,通过图像预处理技术(如图像增强、平滑和锐化)提高图像的质量,便于特征提取和噪声去除。其次,利用CNN对预处理后的图像进行自动缺陷识别。结果表明,该方法在缺陷检测的精度和稳定性方面表现优异。该检测方法不仅能够提升PCB板的质量检测效率和质量控制水平,还可以推广至其他工业生产中的图像检测任务,在工业应用中具有广阔的前景。
With the rapid development of electronic information manufacturing industry,the quality requirements for circuit boards are increasingly high,and traditional manual inspection methods are difficult to meet the high-precision and high-efficiency inspection needs.Therefore,this article proposes a PCB board image defect detection method based on Convolutional Neural Network(CNN).Firstly,image preprocessing techniques such as image enhancement,smoothing,and sharpening are used to improve the quality of the image,facilitating feature extraction and noise removal.Secondly,CNN is used for automatic defect recognition of preprocessed images.The results indicate that this method performs excellently in terms of accuracy and stability in defect detection.This detection method can not only improve the quality inspection efficiency and quality control level of PCB boards,but also be extended to other image inspection tasks in industrial production,with broad prospects in industrial applications.
作者
贾照菊
王献平
程强
冯美艳
JIA Zhaoju;WANG Xianping;CHENG Qiang;FENG Meiyan(School of Aeronautical Engineering,Anyang University,Anyang Henan 455000;Hebi College of Engineering and Technology,Henan Polytechnic University,Hebi Henan 458000)
出处
《软件》
2024年第11期145-147,共3页
Software
关键词
缺陷检测
PCB板
卷积神经网络
电子制造业
defect detection
PCB board
convolutional neural network
electronics manufacturing industry
