YOLO算法在计算机视觉领域中的应用与热点分析

Application and Hotspot Analysis of the YOLO Algorithm in the Field of Computer Vision

摘要: 为了全面分析YOLO算法在我国计算机视觉领域中的实际应用，通过使用CiteSpace计量数据可视化的文献计量法，对取得的计算机视觉领域中的YOLO算法的3741篇论文进行了探究分析。结果表明，我国计算机视觉领域应用YOLO算法的研究机构形成以中国农业大学、中国科学院大学、华南农业大学、西北农林科技大学和南京信息工程大学为核心的五大研究群，我国计算机视觉领域应用YOLO算法的研究人员形成以浙江师范大学朱信忠、西北农林科技大学宋怀波、三峡大学张上、华北理工大学张灿和常熟市中医院王甘红为核心的五大作者群。计算机视觉领域YOLO算法的关键词聚类排名前八的为#0轻量化，#1人脸检测，#2缺陷检测，#3深度学习，#4神经网络，#5 YOLO v5，#6损失函数，#7图像识别。最后通过突变词分析，YOLO算法未来的研究热点将集中在YOLO v5s和激活函数两个关键词上面。

Abstract: To comprehensively analyze the practical application of the YOLO algorithm in China’s computer vision field, this study adopts the bibliometric method with CiteSpace-based quantitative data visualization to explore and analyze 3741 papers on the YOLO algorithm in the computer vision field. The results indicate that, in the field of applying the YOLO algorithm in China’s computer vision sector, research institutions have formed five core research clusters, with China Agricultural University, University of Chinese Academy of Sciences, South China Agricultural University, Northwest A&F University, and Nanjing University of Information Science and Technology as the key hubs. Meanwhile, researchers in this field have established five core author groups, centered on Zhu Xinzhong (Zhejiang Normal University), Song Huaibo (Northwest A&F University), Zhang Shang (China Three Gorges University), Zhang Can (North China University of Science and Technology), and Wang Ganhong (Changshu Traditional Chinese Medicine Hospital). The top 8 keyword clusters of the YOLO algorithm in the computer vision field are #0 Lightweight, #1 Face Detection, #2 Defect Detection, #3 Deep Learning, #4 Neural Network, #5 YOLO v5, #6 Loss Function, and #7 Image Recognition. Finally, through burst word analysis, future research hotspots of the YOLO algorithm will focus on two keywords: YOLO v5s and Activation Function.