基于多时相OLI数据的宁夏大尺度水稻面积遥感估算

发布时间：2018-04-13 15:29

本文选题：遥感 + 作物　；参考：《农业工程学报》2017年15期

【摘要】：为客观获取宁夏水稻面积空间分布信息,也为区域农作物遥感监测奠定技术基础,该文以宁夏回族自治区为研究区域,选择美国LandSat-8携带的陆地成像仪(operational land imager,OLI)数据,采用2016年3月11日-7月01日间的15景影像,基于水稻田耕地与水体特征反射率随着季节变化规律的分析,采用归一化植被指数(normalized difference vegetation index,NDVI)、近红外波段反射率(infrared reflectance,IR)、短波指数(short waved index,SWI)3个指数,以及多时相NDVI最大值、IR最小值、SWI最小值3个衍生指数,共6个指数为基础进行决策分类树构建,对全区水稻进行识别与提取,采用该区水稻面积本底遥感调查结果进行精度验证,水稻种植面积提取误差仅.4.22%,Kappa系数为0.83,水稻空间分布的用户分类精度分别为85.11%,制图精度为81.67%;同时与监督分类方法提取的水稻面积进行对比,该文方法提取水稻的用户精度提高了8.13个百分点,制图精度更是提高了20.01个百分点。研究结果表明,利用中高分辨率的OLI遥感时间序列卫星影像,在大宗农作物时间序列的变化规律分析基础上,构建分类决策树,可以准确地提取大宗农作物种植面积,是区域农作物面积遥感监测业务运行中具有潜力的方法。
[Abstract]:Objective to obtain rice in Ningxia area of spatial information, but also lays the foundation for regional crop remote sensing monitoring, this paper takes the Ningxia Hui Autonomous Region as the study area, land imager choose America carrying LandSat-8 (operational land imager, OLI) data, the March 11, 2016 -7 month 01 day 15 image, paddy cultivated land and water features with reflectance analysis of seasonal variation based on the normalized difference vegetation index (normalized difference vegetation index, NDVI), near infrared reflectance (infrared reflectance, IR (short waved), short wave index index, SWI) 3 index, and multitemporal NDVI maximum IR minimum, the minimum value of SWI 3 derivative index, total the 6 index based decision classification tree, recognition and extraction of the rice, the remote sensing survey results are accuracy test using rice area of the district C, rice planting area extraction error is only.4.22%, the Kappa coefficient is 0.83, the user classification accuracy of spatial distribution of rice were 85.11%, mapping accuracy is 81.67%; at the same time, rice area extraction and supervised classification methods were compared, the extraction method of rice user accuracy increased by 8.13 percentage points, more is to improve the mapping accuracy 20.01 percentage points. The results show that the use of OLI time series remote sensing satellite images with high resolution, the variation in staple crops in time series based on the analysis of the construction of decision tree classification, can accurately extract the planting area of main crops, is a potential method of regional crop area remote sensing monitoring business operation.

【作者单位】：中国农业科学院农业资源与农业区划研究所;
【基金】：国家重点研发计划“粮食作物生长监测诊断与精确栽培技术”课题“作物生长与生产力卫星遥感监测预测”(2016YFD0300603)
【分类号】：S127;S511

【相似文献】