绿洲植被覆盖度遥感信息提取——以敦煌绿洲为例  

Vegetation Cover Information Extraction Technology for Dunhuang Oasis Based on Remote Sensing Images

作　　者：张号[1] 屈建军[1] 张克存[1] 

机构地区：[1]中国科学院寒区旱区环境与工程研究所沙漠与沙漠化重点实验室/敦煌戈壁荒漠生态与环境研究站,甘肃兰州730000

出　　处：《中国沙漠》2015年第2期493-498,共6页Journal of Desert Research

基　　金：国家自然科学基金项目(41371027;41271023)

摘　　要：以敦煌绿洲为研究区,利用Landsat TM遥感数据,通过归一化植被指数(NDVI)和混合像元分解两种方法,提取了敦煌绿洲的植被覆盖度信息。在基于NDVI提取植被覆盖度时,选取了基于NDVI的像元二分模型;在混合像元分解过程中,对遥感影像进行波段反射率归一化处理和最小噪声变换(MNF),确定了3个类型端元:植被、不透水表面/土壤、水体/阴影;最后利用高分辨率遥感影像验证对比了两种提取方法的精度。结果表明:混合像元分解更能准确地提取敦煌地区植被覆盖度信息,其线性相关系数为0.8915,均方根误差为0.0882,而且提取结果更符合实际情况,可以为敦煌植被状况监测及生态环境保护提供科学建议。Taking Dunhuang oasis for example,Normalized Difference Vegetation Index(NDVI)and Spectral Mixture Analysis(SMA)were used to extract the vegetation coverage information from Landsat TM images.While extracting vegetation coverage based on NDVI,we selected a dimidiate pixel model;In the process of SMA,through reflectance normalization and minimum noise transform(MNF),three end members were identified:vegetation,impervious surface/soil,water/shadow.In the end,the results were tested by high resolution Quickbird images.The results showed that SMA could better extract vegetation cover information in arid area with the better accuracy(R=0.8915;RMSE=0.0882).This study could provide the reference for the protection of ecological environment in Dunhuang oasis.

关 键 词：敦煌绿洲 植被覆盖度 归一化植被指数 混合像元分解 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]