踩踏事故预防监控领域科学知识图谱研究*

Research on scientific knowledge map in field of prevention and monitoring on stampede accidents

摘要:

为解决踩踏事故预防监控领域发展现状、发展趋势及发展脉络不清晰问题，运用可视化软件CiteSpace，对Web of Science核心合集及中国知网收集到的365篇相关文献开展知识图谱研究。结果表明：中国在踩踏事故预防监控领域发文量最多，占40.6%，其次为美国，占15.4%，29个国家间仅有25条合作连线；国内外研究热点集中于卷积神经网络等计算机视觉技术在踩踏事故预防监控领域应用；针对“踩踏事故预防监控”领域涵盖学术外延范围，国内外存在一定差别；梳理获得踩踏事故预防领域3条主要发展路径；以深度学习为主的技术研究是当前及未来一段时间研究热点。

Abstract:

In order to solve the problem of unclear development status,development trend and development context in the field of the prevention and monitoring of stampede accidents at home and abroad,the visual software CiteSpace was used to study the knowledge map of 365 related literatures collected from the core collection of Web of Science and CNKI.The results showed that China had the largest number of papers in this field,accounting for 40.6%,followed by the United States with 15.4%,and there were only 25 cooperative connections among 29 countries.Six of the top 20 hot research key words at home and abroad were the same,focusing on the application research of computer vision technology such as convolutional neural network in the field of the prevention and monitoring of stampede accidents.There were some differences in the academic extension scope covered by the field of “prevention and monitoring of stampede accidents” at home and abroad,and the further consensus was needed.Three main development paths of research in this field were obtained.The technical research focusing on deep learning was not only the current research hotspot,but also the research hotspot in the future.