Font Size: a A A

Study On Hyperspectral Characteristics And Leaf Water Content Inversion Of Sand-fixing Vegetation In Horqin Sandy Land

Posted on:2018-05-31Degree:MasterType:Thesis
Country:ChinaCandidate:Y L NiuFull Text:PDF
GTID:2310330518955794Subject:Hydrology and water resources
Abstract/Summary:PDF Full Text Request
The measurement and research of the spectral features of the ground objects was an important part of the remote sensing theory and the basis of the analysis of various remote sensing applications.Because of its high spectral resolution,the ground measured spectral data can accurately reflect the spectral information.The Field Spec4(ASD)portable spectrometer was applied to measure the spectral curves of five typical vegetation(Salix gordejevii,Fringed sagebrush,Ephedra,Artemisia halodendron,Caragana microphylla)in Horqin Sandy land.The spectra were extracted and analyzed.Study on the spectral characteristics of different growth stages and Artemisia halodendron in different coverages.It will provide data support for the inversion and verification of vegetation characteristics and the establishment of spectral database in the future.At the same time,hyperspectral inversion model of Caragana microphylla leaf water content was established.In order to provide the basis to explore the methods and applications of high resolution satellite remote sensing data in the moisture content inversion of sand fixing vegetation in the future.The main results were as follows:1.The spectral form of five typical dune vegetation was roughly the same,but slightly different.Large different bands were 600-690nm and 750-900nm,according to the spectral difference and derivative spectral parameters could distinguish between Fringed sagebrush,Ephedra and Artemisia halodendron,but could not distinguish between Salix gordejevii and Caragana microphylla,2.With vegetation growing season gone on,"green peak position"undergone the process of offset to the red light after the first blue shift.The vegetation spectral "red edge" of Salix gordejevii,Fringed sagebrush,Ephedra and Caragana microphylla had obvious "Twin Peaks",but Artemisia halodendron was not obvious.The "red edge" area of Salix gordejevii,Ephedra and Caragana microphylla appeared "red shift" phenomenon,Salix gordejevii and Fringed sagebrush's "red edge" area at the end of September appeared"blue shift" phenomenon,and the "red edge position" of Artemisia halodendron and Fringed sagebrush showed a slight decreasing trend,the "red shift"phenomenon was not present.Five typical dune vegetation in August and September had the "red edge platform",and Artemisia halodendron,Caragana microphylla on September 27 were sustained in the "red edge platform"position.3.The absorption of five typical dune vegetation and the absorption depth in red band were significantly stronger than that in the blue band.With vegetation growth gone,the change trend of absorption depth of five dune vegetation were same as the absorption area and NDVI.The "red edge","absorption" parameters and timing of NDVI data of different growth stages to identify the effect of vegetation were better than other parameters.When the vegetation classification was carried out in the study area,it was better to choose the multi temporal remote sensing image in the spring and the middle and late autumn.4.Analysis spectral characteristics of Artemisia halodendron in different coverage.And when the coverage was more than 40%,the "red-edge position"and coverage were not directly linked,but the canopy coverage was proportional to the "red edge slope".To analyze continuum removal in the visible band,it increased the absorption depth and decreased the absorption peak area of Artemisia halodendron with coverage increasing.5.In this paper,the wavelet transform were applied to the Caragana microphylla leaf hyperspectral reflectance data of 113 groups in Horqin region.To compared with linear model of Caragana microphylla leaf water content based on wavelet vegetation index inversion and the inversion model based on spectral vegetation indices.The results show that the decision coefficient R2 of multivariate linear model RMCw based on 6 wavelet vegetation index and the RMC estimation model based on3 spectral indices were 0.72 and0.46,the test accuracy R2 were 0.88 and 0.57.
Keywords/Search Tags:Dune vegetation, Spectral characteristics, leaf water content, Wavelet transform, Horqin Sandy Land
PDF Full Text Request
Related items