植被盖度遥感反演模型在稀疏高寒草原的对比研究  

Comparisons on Sparse Alpine Grassland Based on Vegetation Coverage Inversion Models of Remote Sensing

作　　者：夏颖[1,2] 范建容[1] 李磊磊 李炫[1,2] XIA Ying;FAN Jian-rong;LI Lei-lei;LI Xuan(Institute of Mountain Hazards and Environment,CAS,Chengdu 610041;Chinese Academy of Sciences,Beijing 100049;Esri China Information Technology Co.,Ltd,Beijing 100000,China)

机构地区：[1]中国科学院水利部成都山地灾害与环境研究所,成都610041 [2]中国科学院大学,北京100049 [3]易智瑞(中国)信息技术有限公司,北京100000

出　　处：《四川农业大学学报》2017年第1期37-44,59,共9页Journal of Sichuan Agricultural University

基　　金：中国科学院战略性先导科技专项(A类)(XDA05050506)

摘　　要：【目的】基于Landsat8遥感影像数据,反演西藏日喀则地区高寒草原植被盖度。【方法】采用比较常用的3种植被盖度反演模型,即像元二分模型、基于归一化植被指数(NDVI)与土壤调整植被指数(SAVI)的回归模型、改进的三波段梯度差模型,对稀疏高寒草原植被盖度进行反演,并采用照相法实测植被盖度进行反演精度分析。【结果】3种模型对高寒草原植被盖度的反演精度以像元二分模型最高,反演精度为82.02%,其他两种模型均小于80%。【结论】像元二分模型相对于回归模型和改进的三波段梯度差模型更适用于稀疏高寒草原植被盖度的反演。【Objective】The aim of the study was to estimate vegetation coverage of sparse alpinegrassland in Shigatse city based on Landsat8multispectral images.【Method】The dimidiate pixel model,regression model based on NDVI(Normalized Difference Vegetation Index)and SAVI(Soil AdjustedVegetation Index)and modified three-band gradient difference model were used to estimate vegetationcoverage of sparse alpine grassland.The vegetation coverage calculated by photographic method,whichwas used to validate the estimate values.【Results】The dimidiate pixel model had a higher precisioncompared to other two models.its precision was82.02%but the estimation accuracy by other twomodels were less than80%.【Conclusion】The dimidiate pixel model is more suitable for the inversionof vegetation coverage of sparse alpine grassland compared to other models.

关 键 词：植被盖度 稀疏高寒草原 像元二分模型 改进的三波段梯度差模型 回归模型 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]