%0 Journal Article %T 无人机遥感的农作物精细分类研究进展 %T Progress on fine classification of crops based on unmanned aerial vehicle remote sensing %A 田甜,王迪,曾妍,张影,黄青※ %A Tian Tian %A Wang Di %A Zeng Yan %A Zhang Ying %A Huang Qing※ %J 中国农业信息 %J China agricultural informatics %@ 1672-0423 %V 32 %N 2 %D 2020 %P 1-12 %K 无人机;遥感;农作物分类;多光谱;高光谱 %K unmanned aerial vehicle;remote sensing;crop classification;multispectral; hyperspectral %X 【目的】农作物精细分类是面积估算、长势监测、产量预测及灾害评估的重要前提和基 础。近年来,无人机低空遥感技术因其操作成本低、空间分辨率高、灵活性强等优势,成为田 块尺度下农作物精细分类的重要工具。【方法】文章系统总结了国内外近10 余年无人机遥感在 农作物分类领域的研究进展,介绍了目前常用的无人机平台和传感器,归纳了农作物分类特征 及算法的使用情况,指出了无人机遥感农作物精细分类研究存在的问题。【结果】当前无人机 遥感农作物精细分类研究存在一些不足之处:(1)无人机遥感监测面积小,无法在较大尺度区 域实现农作物精准监测。(2)适用于无人机遥感的农作物分类特征仍需进一步挖掘,面向高光 谱影像的农作物分类特征及特征组合尚需进一步明确。(3)分类器使用单一,分类算法的普适性 和稳定性不强。【结论】无人机遥感农作物精细分类研究的发展趋势主要包括3 个方面:(1)无人 机遥感影像与星载遥感数据的高效融合,拓宽无人机的监测范围。(2)面向无人机遥感影像 的农作物分类特征提取与优化研究。(3)适合无人机遥感的农作物分类算法改进。 %X [Purpose]Fine classification of crops is an important prerequisite and basis for area estimation,growth monitoring,yield forecasting and disaster assessment. In recent years, unmanned aerial vehicle(UAV)low-altitude remote sensing technology has become an important tool for fine classification of crops at field scale due to its advantages such as low operating cost, high spatial resolution and strong flexibility.[ Method]This paper systematically summarizes the research progress of UAV remote sensing in fine classification of crops in the past 10 years at home and abroad,introduces the commonly used UAV platforms and sensors,summarizes the selection of crop classification features and the use of classification algorithms,and points out the problems existing in the research.[ Result]There are some shortcomings in the existing research on fine classification of crops by UAV remote sensing:(1)UAV remote sensing monitoring area is small,unable to achieve accurate crop monitoring in large scale areas.( 2)Crop classification features applicable to UAV remote sensing still need to be further explored,and crop classification features and feature combinations oriented to hyperspectral images still need to be further clarified. (3)The single classifier is used,and the universality and stability of the classification algorithm are not strong.[ Conclusion]The development trend of UAV remote sensing crop classification research mainly includes three aspects:(1)Efficient fusion of UAV remote sensing images and satellite-borne remote sensing data will broaden the monitoring range of UAV.( 2)Research on feature extraction and optimization of crop classification for UAV remote sensing images. (3)Development of crop classification algorithms suitable for UAV remote sensing. %R 10.12105/j.issn.1672-0423.20200201 %U http://www.cjarrp.com/zgnyxx/ch/reader/view_abstract.aspx %1 JIS Version 3.0.0