Journal of Integrative Agriculture ›› 2021, Vol. 20 ›› Issue (7): 1944-1957.DOI: 10.1016/S2095-3119(20)63329-9

所属专题: 农业生态环境-遥感合辑Agro-ecosystem & Environment—Romote sensing

• 论文 • 上一篇    下一篇

  

  • 收稿日期:2020-03-08 出版日期:2021-07-01 发布日期:2021-06-02

Monthly composites from Sentinel-1 and Sentinel-2 images for regional major crop mapping with Google Earth Engine

LUO Chong1, LIU Huan-jun1, 2, LU Lü-ping2, LIU Zheng-rong2, KONG Fan-chang2, ZHANG Xin-le2
  

  1. 1 Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun 130102, P.R.China
    2 School of Pubilc Adminstration and Law, Northeast Agricultural University, Harbin 150030, P.R.China
  • Received:2020-03-08 Online:2021-07-01 Published:2021-06-02
  • Contact: Correspondence LIU Huan-jun, E-mail: huanjunliu@yeah.net
  • About author:LUO Chong, E-mail: luochong93@yeah.net;
  • Supported by:
    The paper was funded by the National Key R&D Program of China (2017YFD0201803) and the Talent Recruitment Project of Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences.

摘要:

快速、准确地获取大区域、高分辨率的作物类型分布图对农业精准管理与可持续发展具有重要意义。受遥感影像质量和数据处理能力的限制,使用遥感技术进行大尺度的作物分类仍是一项巨大的挑战。本研究的目的是使用Google Earth Engine(GEE)结合Sentinel-1和Sentinel-2影像绘制黑龙江省的作物分布图,首先收集2018年作物生长关键期(5月至9月)覆盖研究区域所有可用的Sentinel-1与Sentinel-2影像,并对影像进行月度合成,然后将月度合成影像的不同反射率波段、植被指数与极化波段作为输入量结合随机森林方法进行作物分类。结果表明使用本研究提出的方法可以准确地获得黑龙江省作物分布图,作物分类总体精度达到89.75%。本研究还发现相比仅使用传统波段(可见光波段和近红外波段)进行作物分类,增加短波红外波段可以显著改善作物分类的准确性,其次是增加红边波段,增加常见植被指数和Sentinel-1数据对作物分类的精度提升不大。本研究还分析了作物分类的时效性,结果表明当7月份的影像可用时,作物分类精度的提升幅度最大,作物分类的总体精度可以达到80%以上。本研究结果为基于遥感的大尺度、高分辨率作物分布图的制作提供支持。


Abstract:

Rapid and accurate access to large-scale, high-resolution crop-type distribution maps is important for agricultural management and sustainable agricultural development.  Due to the limitations of remote sensing image quality and data processing capabilities, large-scale crop classification is still challenging.  This study aimed to map the distribution of crops in Heilongjiang Province using Google Earth Engine (GEE) and Sentinel-1 and Sentinel-2 images.  We obtained Sentinel-1 and Sentinel-2 images from all the covered study areas in the critical period for crop growth in 2018 (May to September), combined monthly composite images of reflectance bands, vegetation indices and polarization bands as input features, and then performed crop classification using a Random Forest (RF) classifier.  The results show that the Sentinel-1 and Sentinel-2 monthly composite images combined with the RF classifier can accurately generate the crop distribution map of the study area, and the overall accuracy (OA) reached 89.75%.  Through experiments, we also found that the classification performance using time-series images is significantly better than that using single-period images.  Compared with the use of traditional bands only (i.e., the visible and near-infrared bands), the addition of shortwave infrared bands can improve the accuracy of crop classification most significantly, followed by the addition of red-edge bands.  Adding common vegetation indices and Sentinel-1 data to the crop classification improved the overall classification accuracy and the OA by 0.2 and 0.6%, respectively, compared to using only the Sentinel-2 reflectance bands.  The analysis of timeliness revealed that when the July image is available, the increase in the accuracy of crop classification is the highest.  When the Sentinel-1 and Sentinel-2 images for May, June, and July are available, an OA greater than 80% can be achieved.  The results of this study are applicable to large-scale, high-resolution crop classification and provide key technologies for remote sensing-based crop classification in small-scale agricultural areas.

Key words: Sentinel-1 ,  Sentinel-2 ,  monthly composites ,  crop mapping ,  Google Earth Engine