Abstract:

In order to systematically investigate and intuitively reflect the status quo, hotspots and frontiers of machine vision defect detection research in China, 588 related literatures from 1995 to 2020 are selected based on CNKI database for bibliometric measurement and knowledge map analysis. From 1995 to 2020, the research on machine vision defect detection in China has generally experienced three stages：preliminary exploration（1995-2005）, rapid growth（2006-2011） and high-speed growth（2012-2020）. Among the research keywords, machine vision, image processing, defect detection and support vector machine are the core words in this field. In keyword evolution analysis, the research is expanded from defect detection, image processing and deep learning focusing on machine vision to the subdivision fields such as surface defects, image segmentation and digital image processing. And WANG Yaonan, PENG Yu, FAN Tao and their team members are among the most prolific authors. The School of Automation of Guangdong University of Technology and the School of Manufacturing Science and Engineering of Sichuan University, plays important roles in academic research. In regard to journal carriers, a core group is formed of nine journals, which includs Progress [in] Laser [and] Optoelectronics  and Manufacturing Automation.

Key words: machine vision, defect detection, bibliometrics measurement, knowledge mapping, CiteSpace

摘要：

为系统考察和直观反映中国机器视觉缺陷检测研究的现状、热点和前沿，以“中国机器视觉缺陷检测研究”为主题，基于中国知网（CNKI）数据库，筛选出1995—2020年相关文献共588篇进行文献计量和知识图谱分析。结果表明：1995—2020年，中国机器视觉缺陷检测研究大体经历了初步探索（1995—2005年）、快速增长（2006—2011年）和高速增长（2012—2020年）三个阶段。在研究关键词中，机器视觉、图像处理、缺陷检测、支持向量机等为该领域的核心词。在关键词演进分析中，该领域研究热点总体上从围绕机器视觉的缺陷检测、图像处理、深度学习逐步向表面缺陷、图像分割、数字图像处理等细分领域拓展。在发文作者上，以王耀南、彭玉、范涛及其团队为引领而取得的成果最为丰硕。在研究机构上，以广东工业大学自动化学院、四川大学制造科学与工程学院为代表的科研院所占据重要地位。在刊文载体上，已形成由《激光与光电子学进展》《制造业自动化》等9种期刊组成的核心期刊群。

关键词: 机器视觉, 缺陷检测, 文献计量, 知识图谱, CiteSpace