基于MODIS的松花江流域主要作物分布提取  

Classification of major crops using MODIS data in the Songhua River Basin

作　　者：冯颖 郭英[2] 陈晓璐 刘孟竹 沈彦俊[2,3] FENG Ying;GUO Ying;CHEN Xiaolu;LIU Mengzhu;SHEN Yanjun(School of Geographical Sciences,Hebei Normal University,Shijiazhuang 050024,China;Center for Agricultural Resources Research,Institute of Genetics and Developmental Biology,Chinese Academy of Sciences/Key Laboratory of Agricultural Water Resources,Chinese Academy of Sciences,Shijiazhuang 050022,China;School of Advanced Agricultural Sciences,University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]河北师范大学地理科学学院,石家庄050024 [2]中国科学院遗传与发育生物学研究所农业资源研究中心/中国科学院农业水资源重点实验室,石家庄050022 [3]中国科学院大学现代农业科学学院,北京100049

出　　处：《中国生态农业学报（中英文）》2023年第10期1602-1612,共11页Chinese Journal of Eco-Agriculture

基　　金：中国科学院战略性先导科技专项子课题(XDA28020503,XDA28020500);中国科学院农业水资源重点实验室重点培养项目(ZDKT201801)资助。

摘　　要：准确地掌握并监测区域内的作物分布及时空变化信息对科学指导流域内农业生产布局具有重要指导意义。本研究选择中国的粮食主产区——松花江流域为研究区,基于MODIS数据计算的归一化植被指数(NDVI)、植被增强型指数(EVI)及地表水分指数(LSWI)指标,结合不同作物生长发育的物候期构建决策树模型,对2000年和2020年研究区内主要作物——水稻、玉米和大豆种植区进行了提取;提取结果通过实地考察、Google Earth选点和统计年鉴数据进行了验证。水稻、玉米和大豆提取精度分别为0.9090、0.9026和0.8200,Kappa系数为0.79,总体精度为0.84。松花江流域种植农作物类型以玉米为主,水稻、大豆为辅,流域内形成“北大豆,南玉米,河流附近皆水稻”的生产格局,是我国重要的玉米、水稻和大豆生产基地。2000—2020年流域种植规模处于持续扩张状态,作物种植结构显著改变,总种植面积由95556.26 km^(2)增加到173070.00 km^(2),增幅为81.12%,其中水稻、玉米和大豆种植面积分别增加24911.36 km^(2)、54432.07 km^(2)、20719.77 km^(2),水稻、玉米和大豆种植面积占总种植面积比例分别增加8.56%、16.86%和0.39%。从空间变化来看,水稻种植面积在松花江流域范围内大部分区域明显增加,尤其是在河流附近水资源丰富的区域;玉米种植面积在流域内大部分区域显著增加;大豆种植面积在流域北部偏东区域增加明显。松花江流域各市形成了具有特色的种植结构,绝大部分地区由双作物主导型转变为玉米主导型;2020年流域内新增水稻主导型市,玉米-水稻主导型市和大豆-水稻主导型市消失,玉米-水稻主导型市和玉米-大豆主导型市均转变为玉米主导型市,作物种植类型逐渐趋于集中。本文提取的作物分布结果与松花江流域内实际情况一致,可为松花江流域的种植结构调整及指导农业生产提供科学支撑。In 2000,China launched a series of positive policies to promote agricultural development in Northeast China,thus causing a rapid expansion of the cropping scale and a change of cropping structure in the Songhua River Basin.It is important to reveal the changes in cropping structure in the Songhua River Basin to improve its future supply capacity,ensure national food security,and achieve the adjustment of food production.In this study,we selected the Songhua River Basin,a major grain-producing area in China,as the study area.Based on the normalized difference vegetation index(NDVI),enhanced vegetation index(EVI),and land surface water index(LSWI)from MODIS data,a decision tree model was constructed by combining the phenological periods of different crops.The results of crop extraction were verified using observed datasets,verification points selected in Google Earth,and statistical almanac data.The values of extraction accuracy for rice,maize,and soybean were 0.9090,0.9026,and 0.8200,respectively,with a Kappa coefficient of 0.79 and an overall accuracy of 0.84.The crop types in the Songhua River Basin were dominated by maize and supplemented by rice and soybean,forming a crop pattern of“soybean in the north,maize in the south,and rice near rivers”.The entire crop planting area of the Songhua River Basin was in a state of continuous expansion from 2000 to 2020,with the total planting area increasing from 95556.26 km^(2)to 173070.00 km^(2),an increase of 81.12%.The planting area of rice,maize,and soybean increased by 24911.36 km^(2),54432.07 km^(2),and 20719.77 km^(2),respectively.Overall,the proportion of the rice planting area to the total planting area increased by 8.56%,and those of maize and soybean increased by 16.86%and 0.39%,respectively.The planting areas of rice and maize increased significantly in most areas of the Songhua River Basin,and the rice expanded mainly in the areas near rivers.The planting area of soybean increased significantly in the eastern part of the northern part of the basin.The citi

关 键 词：松花江流域 作物提取 MODIS 粮食安全 植被指数 水稻 玉米 大豆 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]