Remote Sensing Estimation Of Vegetation Above-Ground Biomass Based On Multi-Source Data

Based on the field survey data and multi-source remote sensing data,the above-ground biomass of vegetation in the Weigan-Kuqa River Delta Oasis was accurately estimated by combining the multivariate statistical analysis method with geographic information space technology.According to GF2 high-resolution remote sensing image data and field survey sample plot data,vegetation information is classified and extracted,and a quantitative remote sensing estimation model of vegetation aboveground biomass is constructed in an experimental plot.Using Landsat 8 OLI image data to extract vegetation index and band combination factors,combining with the estimation model results of previous experiments for comparative analysis,a large-scale optimal inversion model of vegetation aboveground biomass is established,thus providing theoretical reference for revealing the spatial distribution of vegetation aboveground biomass in Weiku Oasis,mastering the stability of desert oasis ecosystem in arid areas,and regional land desertification control and carbon storage research.The results are as follows:(1)Two methods of supervised classification and object-oriented are used to extract ground object information from GF2 image of experimental district.The results of post-classification processing show that the overall classification accuracy of the object-oriented classification method is higher than that of the supervised classification.Finally,the object-oriented classification method is selected to extract the vegetation information in the study area.The overall classification accuracy is96.64%,and the kappa coefficient is 0.95.(2)By combining the field data with contemporaneous remote sensing image data in the same period,the inversion model of vegetation above-ground biomass in the study area could get preferable inversion results.The reverse performance of the tree aboveground biomass optimal model is an MSR model with B253 as the independent variable,while the shrub aboveground biomass optimal model is an MSR model with NDVI and EVI as the independent variables.The optimal model of herbaceous aboveground biomass is a quadratic model with RVI as the modeling independent variable.PLSR model is the best inversion model of crop aboveground biomass.RMSE and MAE values of the optimal inversion model for aboveground biomass of various vegetation types are within a reasonable range,and the regression test of the optimal inversion model is extremely significant at the level of 0.001.(3)According to the optimal inversion model of vegetation aboveground biomass in the study area,it is estimated by the inversion model that the total aboveground biomass of vegetation in the study area is 44.926×10⁸kg.Shrubs,the highest total above-ground biomass is crop above-ground biomass;the average above-ground biomass of vegetation is crops(0.603kg·m^-2),arbor(0.584kg·m^-2),shrubs(0.406kg·m^-2)and herbs(0.236kg·m^-2).In summary,in this paper,the above-ground biomass estimation of representative vegetation in Wei-Ku oasis is carried out.By constructing the inversion model,it could fully reflect the spatial distribution of the above-ground biomass of the vegetation in Wei-Ku oasis,which is of great practical significance for the accurate estimation of the above-ground biomass of the vegetation in Wei-Ku oasis,and also provides theoretical reference for the further research on vegetation above-ground biomass estimation in other areas.