基于CiteSpace的小目标检测研究热点与发展趋势探析

Research Hotspots and Development Trends of Small Target Detection Based on CiteSpace

摘要: 为系统梳理小目标检测领域的研究脉络与发展态势，精准捕捉领域知识结构与前沿动态，以中国知网收录的1728篇核心期刊论文为研究对象，综合运用文献计量法与CiteSpace可视化分析技术，从文献时间分布、关键词共现网络、聚类结构及突现特征四个维度展开深度分析。研究结果表明，小目标检测领域的发展历程呈现萌芽探索、平稳积累、爆发增长三阶段演化特征；领域研究热点已形成算法优化、场景应用、技术支撑相互关联的三层体系；前沿演进路径遵循从传统图像处理方法到深度学习深度融合，再到轻量化模型工程化落地的清晰逻辑。本文通过量化分析与质性研究相结合的方式，全面呈现小目标检测领域的学术图景，可为科研人员把握领域研究动态、布局创新研究方向提供重要参考，同时为相关技术的产业化落地提供理论支撑与实践指导。

Abstract: To systematically sort out the research context and development trend in the field of small target detection, and accurately capture the domain knowledge structure and cutting-edge dynamics, this paper takes 1728 core journal papers included in China National Knowledge Infrastructure (CNKI) as the research objects, and conducts an in-depth analysis from four dimensions: temporal distribution of literature, keyword co-occurrence network, clustering structure and burst characteristics, by comprehensively adopting bibliometrics and CiteSpace visual analysis technology. The research results show that the development process of the small target detection field presents distinct evolutionary characteristics of three stages, namely embryonic exploration, steady accumulation and explosive growth; the research hotspots in this field have formed an interrelated three-tier system composed of algorithm optimization, scenario application and technical support; the cutting-edge evolution path follows a clear logic from traditional image processing methods to the in-depth integration of deep learning, and then to the engineering implementation of lightweight models. By combining quantitative analysis with qualitative research, this paper comprehensively presents the academic landscape of the small target detection field, which can provide an important reference for researchers to grasp the research dynamics and layout innovative research directions in this field, and also offer theoretical support and practical guidance for the industrialization of relevant technologies.