Landsat-7 ETM+归一化差分植被指数影像真实性检验  

Validation of Landsat-7 ETM+NDVI Imagery

作　　者：栾海军[1,2] 章欣欣[1] 田庆久[2,3] 朱晓玲[1] 聂芹[1] 

机构地区：[1]厦门理工学院计算机与信息工程学院,厦门361024 [2]南京大学国际地球系统科学研究所,南京210093 [3]江苏省地理信息技术重点实验室,南京210093

出　　处：《遥感信息》2016年第6期138-146,共9页Remote Sensing Information

基　　金：国家自然科学基金(41601350;41501447);福建省测绘地理信息局局校合作项目(2015JX04);福建省中青年教师教育科研项目(A类)(JA14231)

摘　　要：作为一种常用的地表参数产品,ETM+NDVI的真实性检验具有重要意义,但是目前对其研究较少。本文提出结合GeoEye-1高分辨率遥感影像及其NDVI尺度转换模型的方法进行该产品的真实性检验。基于分形理论及5指标评价体系(r>=0.8。ETM＋NDVI is a frequently used land surface parameter. Its validation is rather valuable, whereas the relevant research is little. Therefore, a validation methodology of ETM＋NDVI was proposed based on the combination of GeoEye 1 high resolution image and its NDVI scaling model in our research. For the experimental GeoEye 1 image, NDVI＇s CSSM （continuous spatial scaling model） was established on the basis of fractal theory and five-index estimation system （r〉=0.8、p〈0.05、rlo〉=r〈=rup and Max_of_abs（Error） 〈 = 0.05）. Meanwhile,the spectral normalization was made for GeoEye-1 and ETM＋ experimental images. Then the validation of ETM＋ NDVI image was implemented. For plenty of ETM＋ sample images,the experimental results manifested that：because of the serious influence of ETM＋ images＇ bad stripes and the limited function of destriping operation, NDVI of the sample images had rather big errors, and for the entire ETM＋ experimental image the error was about 25%. Hence the ETM ＋ NDVI product was not suitable for being applied directly. In the meantime, the potential value of NDVI＇s CSSM in validation had been demonstrated.

关 键 词：NDVI 真实性检验 ETM+ 空间尺度转换 分形理论 GeoEye-1 

分 类 号：TP701[自动化与计算机技术—检测技术与自动化装置]