基于文献计量的农作物种植结构遥感提取发展态势分析

Development situation analysis of remote sensing extraction of crop planting structure based on bibliometrics

及时获取农作物种植结构对粮食安全及农业可持续发展具有重要意义。为了解当前农作物种植结构遥感提取研究前沿和进展,利用文献计量法探究了该领域近20年的研究现状及热点。利用Vosviewer软件和Bibliometrix程序包,分析了Web of Science核心合集和中国知网中2000-2022年农作物种植结构遥感提取领域的867篇文献,从发文量、发文国家、国内发文机构、载文期刊、关键词共现及关键词时序等角度进行可视化分析。农作物种植结构遥感提取研究领域整体发文量呈指数上升;英文发文量最多的国家是中国,为197篇,国际合作最多的也是中国,中文发文量较多的机构是中国科学院(89篇)、中国农业科学院(38篇);研究主要集中于如何选择遥感影像、提取方法以及特征;通过对热点关键词进行时序变化分析发现,当前和未来的研究热点数据是基于时间序列的多源数据融合,热门分类方法主要是以面向对象并结合随机森林、卷积神经网络进行提取,热门特征指数还是以归一化植被指数(NDVI)为主并结合其他特征选择最优特征组合。农作物种植结构遥感提取研究未来将聚焦在如何向低成本、高效率、高精度提取发展。

It is of great significance to acquire crop planting structure timely for food security and sustainable agricultural development.The research status and focuses in the field in recent 20 years were explored by bibliometric analysis to learn about the research frontier and progress on crop planting structure mapping by remote sensing extraction.Literatures with a number of 867 in the field of crop planting structure mapping by remote sensing extraction in the core collection of Web of Science and CNKI from 2000 to 2022 were analyzed.Visualized analysis was made by Vosviewer software and Bibliometrix program package,from the perspectives of published articles number,countries of publication,domestic institutions of publication,journals of publication,keywords co-occurrence and keywords timing sequence.The total publication number in research field of crop planting structure mapping by remote sensing extraction had increased exponentially.China was the country with the largest number of articles published in English,which was 197.Besides,China was the country with the greatest contribution to international cooperation.The institutions with large number of articles published in Chinese were Chinese Academy of Sciences and Chinese Academy of Agricultural Sciences,with articles number of 89 and 38,respectively.The articles mainly focused on selection of remote sensing images,extraction methods and features.It was found through analysis on time sequence changes of hot keywords that,the hot data of current and future researches were multi-source data fusion based on time series,the hot classification method was mainly led by the object-oriented mapping in combination with random forest and convolutional neural networks,and the hot feature indexes were still dominated by the normalized difference vegetation index(NDVI)combined with other features to select the optimal combination of the features.In conclusion,the research on the crop planting structure mapping by remote sensing extraction will focus on extraction with low-cost,high-efficiency and high-precision.