Study on Hyperspectral Characteristics of Biological Soil Crust in Water and Wind Erosion Crisscross Region of the Loess PlateauCN
Abstract：The study of hyperspectral characteristics of biological soil crust in the water-wind-erosion crisscross region of the loess plateau has laid a good theoretical foundation and technical support for the remote sensing identification and corrosion resistance evaluation of large-scale biological soil crust.Project using a spectral measurement technology and hyperspectral remote sensing technology,water erosion wind erosion in the loess plateau staggered transition zone as the research area,based on shenmu field ecological experimental station of base,carried out in different types of biological soil crust spectrum characteristics analysis,biological soil crust change water characteristics influence on spectral characteristics,and based on the unmanned aerial vehicle(UAV)hyperspectral remote sensing technology of biological soil crust recognition research.The following research results have been obtained:(1)the algal biogenic soil crust and soil have similar spectral characteristics,and the spectral curve has no obvious "peak-valley" characteristics;The spectrum characteristics of algal biological soil crust mainly show the change rule of spectral reflectance decreasing with the increase of biological soil crust coverage.(2)the spectra of soil crust of mosses show similar characteristics to those of higher plants,forming reflection peaks in the green band and absorption valleys in the red band,as well as high reflection in the near-infrared band;But at 760~930 nm,the spectral Slope of moss biogenic soil crust was significantly higher than that of higher plants,and the spectral Slope(930/760)was 2.5-4.5 times higher than that of higher plants.(3)the crust of bare land,algae and mosses showed the spectral characteristics of decreased reflectivity and deepened water absorption valley in the humid environment.The hyperspectral index(RVI,DVI,NDVI,SAVI,MSAVI and OSAVI)of bare soil and algal soil crust under dry conditions was higher than that under wet conditions.(4)with the increase of mulch degree,moss biogenic soil crust shows obvious "blue shift","yellow shift" and "red shift".Multivariate linear regression model with different coverage of biological soil crust coverage estimation has better fitting results,the linear spectral mixture model for mosses of biological soil crust classification result is better than that of soil(5)based on unmanned aerial vehicle(uav)hyperspectral image using supervised classification,spectral matching technique,six kinds of hyperspectral indices,which can identify the biological soil crust in the studied area,the maximum likelihood method,the Pope to envelope information(SID),chlorophyll absorption index(CARI)the highest recognition accuracy.This study canprovide strong support for the extraction and remote sensing monitoring of biological soil crust on a large scale,and has important scientific value and practical significance for elucidating the role of biological soil crust in a specific ecosystem.