|
|
|
|
  • 软件名称:基于谐波分析和线性光谱模型的耕地信息提取
  • 软件大小: 0.00 B
  • 软件评级: ★★★
  • 开 发 商: 周玉洁,王卷乐,郭海会
  • 软件来源: 《遥感技术与应用》
  • 解压密码:www.gissky.net

资源简介

摘要: 耕地是重要的农业资源,如何利用遥感技术快速准确地提取耕地信息是目前研究的热点。利用2000年MODIS/EVI时间序列数据提取关中地区耕地资源信息。以不同地类的EVI时间序列数据年内变化差异为分类依据,采用时间序列谐波分析法对全年时间谱EVI数据进行重构分析,减少噪音对信息提取的影响。经最小噪声分离变换(MNF变换)、纯净像元指数(PPI)计算以及N维可视化工具进行人机交互选取植被、耕地、城镇和水体4种端元,基于线性光谱混合模型,获取该地区耕地资源分布信息。通过与同年1∶10万土地利用数据对比验证,本研究提取的耕地总体精度为83%。研究表明:基于时间序列谐波分析法对EVI数据进行重构,利用不同地类的特征差异,采用混合像元分解的方法,可以精确获取耕地资源定量信息。该方法可为长期、大范围、动态的耕地分布和变化遥感监测提供技术参考,同时为国土资源管理部门提供决策支持。 关键词: 耕地资源;  MODIS/EVI;  谐波分析;  线性光谱混合模型(LSMM);  关中地区     Abstract: Cultivated land is an important agricultural resources and how to extract information quickly and accurately using remote sensing technology is an research hotspot currently.In this paper,time series of MODIS/EVI data was used to extract cultivated resource in GuanZhong during 2000.According to the classification basis,the different land type exhibiting distinctive seasonal patterns of EVI variation have strong periodic characteristic.Harmonic Analysis of Time Series was applied to EVI time\|series data in order to minimize the influence of noise on information extraction.After performed MNF transformation and PPI algorithm,four endmembers are selected including vegetation,cultivated land,urban and water using n\|Dimensional Visualizer Tool of ENVI software.And then,cultivated land was estimated based on a Linear Spectral Mixture Model (LSMM).Through comparison validation of the 1∶10 000 land use data of the study area.This study finds that the total accuracy of cultivated land is 83%.The results indicated that high accuracy of cultivated land of quantitative information is available based on mixed pixel decomposition according to the characteristic difference of different land type,as well as EVI time\|series data reconstruction using the Harmonic Analysis of Time Series.This method can provide technical reference for long\|term,large\|scale,dynamic remote sensing monitoring of the distribution and change of cultivated land and  

下载说明

·如果您发现该资源不能下载,请通知管理员.gissky@gmail.com

·为确保下载的资源能正常使用,请使用[WinRAR v3.8]或以上版本解压本站资源,缺省解压密码www.gissky.net ,如果是压缩文件为分卷多文件,请依次下载每一个文件,并按照顺序命名为1.rar,2.rar,3.rar...,然后鼠标右击1.rar解压.

·为了保证您快速的下载速度,我们推荐您使用[网际快车]等专业工具下载.

·站内提供的资源纯属学习交流之用,如侵犯您的版权请与我们联系.