计算机科学与探索

• 学术研究 •    下一篇

基于机器视觉的PCB缺陷检测算法研究综述

杨思念, 曹立佳, 杨洋, 郭川东   

  1. 1. 四川轻化工大学 计算机科学与工程学院, 四川 宜宾 644000
    2. 人工智能四川省重点实验室, 四川 宜宾 644000
    3. 企业信息化与物联网测控技术四川省高校重点实验室, 四川 宜宾 644000
    4. 四川轻化工大学 自动化与信息工程学院, 四川 宜宾 644000

Review of PCB Defect Detection Algorithm Based on Machine Vision

YANG Sinian,  CAO Lijia,  YANG Yang,  GUO Chuandong   

  1. 1. School of Computing Science and Engineering, Sichuan University of Science & Engineering, Yibin, Sichuan 644000, China
    2. Artificial Intelligence Key Laboratory of Sichuan Province, Yibin, Sichuan 644000, China
    3. Key Laboratory of Higher Education of Sichuan Province for Enterprise Informationalization and Internet of Things, Yibin, Sichuan 644000, China
    4. School of Automation and Information Engineering, Sichuan University of Science & Engineering, Yibin, Sichuan 644000, China

摘要: 印刷电路板(Printed Circuit Board, PCB)作为电子产品的核心组成部分,其质量直接影响产品的可靠性。随着电子产品朝着更轻、更薄、更精密的方向发展,基于机器视觉的PCB缺陷检测面临诸如微小缺陷难以检测等挑战。为深入研究PCB缺陷检测技术,根据其发展历程对各阶段的算法进行了详细探讨。首先,指出了该领域面临的主要挑战,并介绍了传统PCB缺陷检测方法及其局限性。接着,从传统机器学习和深度学习两个角度系统回顾了近几年PCB缺陷检测所采用的方法及其优缺点。随后,对PCB缺陷检测算法常用的评价指标和主流数据集进行了归纳,并对近三年在PCB Defect、DeepPCB和HRIPCB三个数据集上的最新研究方法进行了性能比较,分析了产生差异化的原因。最后,基于当前现状和亟待解决的问题,对未来的发展趋势进行了展望,旨在为后续相关研究提供参考。

关键词: PCB, 缺陷检测, 机器视觉, 机器学习, 深度学习

Abstract: Printed Circuit Board (PCB) as a core component of electronic products, its quality directly affects the reliability of the product. As electronics move toward lighter, thinner, and more sophisticated products, machine vision-based PCB defect detection faces challenges such as the difficulty of detecting tiny defects. In order to further study the PCB defect detection technology, the algorithms of each stage are discussed in detail according to the development history. First, the main challenges in the field are pointed out, and traditional PCB defect detection methods and their limitations are introduced. Then, from the perspective of traditional machine learning and deep learning, we systematically reviewed the PCB defect detection methods and their advantages and disadvantages in recent years. Then, we summarize the commonly used evaluation indicators and mainstream data sets of PCB Defect detection algorithms, and compare the performance of the latest research methods on PCB defect, DeepPCB and HRIPCB data sets in the past three years, and analyze the reasons for the differences. Finally, based on the current situation and the problems to be solved, the future development trend is prospected, aiming at providing reference for the subsequent relevant research.

Key words: PCB, defect detection, machine vision, machine learning, deep learning