Vegetation Indices Based On Multi-source Remote Sensing Data Of The Qinghai Lake Basin

In this study, within the scope of the Qinghai Lake basin vegetation for the study,based on a variety of remote sensing data, remote sensing image processing andanalysis, extraction portion vegetation index, combined with field sampling dataobtained by biomass, analysis of different remote sensing images and related biomassover the same period sex, type of suitable vegetation index analysis Qinghai Lakebasin, providing advice and theoretical basis for the study of vegetation index andvegetation protection Qinghai Lake basin based on the results.(1) Select MODIS250m, Landsat5TM and Landsat8OIL30m and Rapideye5mremote sensing images, perform radiometric calibration, use dark element method andFLAASH models like MODTRAN radiative transfer model based on the imagesatmospheric correction, transformation to extract vegetation index for surface albedo,and field measurements of biological the amount of correlation analysis, thecorrelation coefficient;(2) Based on field data fitting soil bare soil spectral line equation, the equation ofthe soil line slope and intercept were1.1055and0.0422, into vegetation index valuesrelating to soil vegetation index calculation and expression lines;(3) Correlation was statistically significant correlation between grass biomassand SAVI highest correlation, correlation coefficient0.565MODIS data of vegetationindex and measured biomass; then followed DVI, PVI, MSAVI, RVI, OSAVI,TSAVI,NDVI,coefficientTo0.558,0.557,0.555,0.549,0.546,0.542,0.494; Landsat5TMbiomass and measured data OSAVI highest correlation coefficient reached0.605;then followed MSAVI, SAVI, NDVI, TSAVI, RVI, DVI, PVI, EVI and GEMI,0.604,0.603,0.601,0.598,0.598,0.591,0.584,0.572and correlation coefficients were0.519; extracted10images are planted on Landsat8OLI Index (GEMI does not passinspection and EVI0.05significant levels significantly correlated), RVI highestcorrelation, reached0.649, DVI and MSAVI second, respectively,0.637and0.612;SAVI, OSAVI, PVI and NDVI correlation coefficients were between0.5-0.6, respectively0.595,0.574,0.555and0.525, TSAVI lowest, only0.458; Based Rapideyeimages, RVI highest correlation, reaching0.594; followed SAVI, reaching0.569, thenfollowed: TSAVI, NDVI, OSAVI, MSAVI, PVI, EVI, DVI and GEMI, correlationcoefficients were:0.568,0.565,0.561,0.559,0.549,0.523and0.438, two-tailed valueof0.000(GEMI0.001);(4) The ability to detect vegetation details SAVI, RVI, PVI, DVI, OSAVI,MSAVI and GEMI Landsat8OIL data ranges from the largest, TSAVI, EVI andNDVI images are Landsat5TM maximum ranges, indicating that the current probe isstill the Landsat series of satellites vegetation index video is the best.(5) According to the measured spectral data obtained five types of QinghaiLake basin vegetation red edge position temperate steppe719.489nm, mountainmeadow720.419nm, lowland meadow720.379nm, desert steppe and alpine meadow718.695nm720.312nm.