基于人工智能的PCB缺陷检测系统  

PCB Detection System Based On Artificial Intelligence

作　　者：董大兴[1] 刘捍植 武泽 樊永清 DONG Da-xing;LIU Han-zhi;WU Ze;FAN Yong-qing(College of Science of Nanjing University of Aeronautics and Astronautics,Nanjing 210000 China)

机构地区：[1]南京航空航天大学理学院,江苏南京210000

出　　处：《自动化技术与应用》2023年第3期129-133,共5页Techniques of Automation and Applications

摘　　要：PCB缺陷检测在现代电子信息技术中越来越重要,因此提高检测的效率和准确性是亟需研究的一个问题。针对这一问题,设计了一套基于人工智能识别的PCB故障智能检测系统。该系统首先利用硬件系统采集图像,再通过软件系统完成灰度化、中值滤波等图像预处理操作。之后利用Canny算子和霍夫变换获取边缘信息,并分割元器件图像。最后,利用事先训练好的神经网络模型实现故障识别。实际测试表明,该检测系统的准确率大于90%,因此该系统在PCB质量检测领域具有实际应用的意义。PCB defect detection is becoming more and more important in modern electronic information technology,therefore,improving the efficiency and accuracy of detection is an issue that needs urgent research.An artificial intelligence detection system with the PCB defect as detection object is proposed.The system first uses the hardware system to collect images,and then completes image preprocessing operations such as graying,median filtering through the software parts.After that,canny operator and hough transform are used to obtain the edge information and segment the component image.Finally,the neural network model trained in advance is used to realize fault recognition.The practical test shows that the accuracy of the detection system is more than 90%.Therefore,the system has practical application in the field of PCB quality detection.

关 键 词：PCB 缺陷检测 图像处理 神经网络 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]