| The scientific and effective evaluation of forest conditions and forest management cannot be separated from the systematic,comprehensive,accurate,timely and reliable investigation of various forest parameters.The investigation data provides theoretical basis and reference for scientific and reasonable decision-making of forest management and management,which is of great significance for sustainable forest management and sustainable development of forestry.Although the traditional forest survey can obtain accurate data,it requires huge manpower,material resources and financial resources,usually takes a long time,and the survey results are subjective.However,the development of remote sensing technology provides a new means for it.Taking West Hunan as the research object,this study was based on the data from the 78 plots in 2007 and the remote sensing image from SPOT-5 in the same period,and used the full subset regression model to build the estimation model of various stand parameters,which mainly includes the traditional forest structure parameters(NT,BA,SV),forest diversity parameters(SDDBH,GC,QMD,Shannon index,Simpson index and Pielou index),the stand spatial structure parameters(W,U,M),forest ecological function index(FEFI),and arbor carbon on the ground.The box-cox transform was used to eliminate the heterosceasticity of data,and the left-one cross-validation method was used to verify the accuracy of the above model,as well as the normality and homogeneity of variance of the residuals of SW test and NCV test.Results show that using the SPOT-5 satellite images to extract the texture and spectral information can reliably predict BA,SV,W,U,M,QMD,Shannon index,Simpson index and Pielou index,FEFI and the arbor carbon on the ground,Radj2 can reach more than 0.5,and the model of the distribution of the residual homogeneously near zero,no obvious change trend.The established model can predict stand variables,provide data support for rapid,economic and quantitative evaluation of stand variables,and provide theoretical support for effective forest management and decision-making. |
