Identification and mapping of soybean and maize crops based on Sentinel-2 data

作     者：Bao She Yuying Yang Zhigen Zhao Linsheng Huang Dong Liang Dongyan Zhang 

作者机构：School of GeomaticsAnhui University of Science&TechnologyHuainan 232001AnhuiChina National Engineering Research Center for Agro-Ecological Big Data Analysis&ApplicationAnhui UniversityHefei 230601China 

出 版 物：《International Journal of Agricultural and Biological Engineering》 (国际农业与生物工程学报（英文）)

年 卷 期：2020年第13卷第6期

页      面：171-182页

核心收录：

学科分类：0202[经济学-应用经济学] 02[经济学] 020205[经济学-产业经济学] 

基　　金：This work was supported by the Key Project of Natural Science Research of Education Department of Anhui Province(Grant No.KJ2019A0120) the National Key Research and Development Program of China(2019YFE0115200) and the Open Research Fund of National Engineering Research Center for Agro-Ecological Big Data Analysis&Application(Grant No.AE2018011) The authors appreciate the European Space Agency(ESA)which provides Sentinel-2 images for this research the editor and anonymous reviewers for their valuable comments that helped us to improve this manuscript 

主　　题：soybean and maize crop identification Sentinel-2 data machine learning feature selection 

摘      要：Soybean and maize are important raw materials for the production of food and livestock *** mapping of these two crops is of great significance to crop management,yield estimation,and crop-damage *** this study,two towns in Guoyang County,Anhui Province,China,were selected as the study area,and Sentinel-2 images were adopted to map the distributions of both crops in the 2019 growing *** data obtained on August 18(early pod-setting stage of soybean)was determined to be the most applicable to soybean and maize mapping by means of the Jeffries-Matusita(JM)***,three machine-learning algorithms,i.e.,random forest(RF),support vector machine(SVM)and back-propagation neural network(BPNN)were employed and their respective performance in crop identification was evaluated with the aid of 254 ground truth *** appeared that RF with a Kappa of 0.83 was superior to the other two ***,twenty candidate features containing the reflectance of ten spectral bands(spatial resolution at 10 m or 20 m)and ten remote-sensing indices were input into the RF algorithm to conduct an important *** features were screened out and served as the optimum subset,the mapping results of which were assessed based on the ground truth derived from the unmanned aerial vehicle(UAV)images covering six ground *** optimum feature-subset achieved high-accuracy crop mapping,with a reduction of data volume by 65%compared with the total twenty features,which also overrode the performance of ten spectral ***,feature-optimization had great potential in the identification of the two ***,the findings of this study can provide a valuable reference for mapping soybean and maize in areas with a fragmented landscape of farmland and complex planting structure.