Remote Sensing Estimation of Pinus yunnanensis Accumulation Based on GS-RF Optimization Algorithm—Taking Zhanyi District of Qujing City as an Example
Accurate estimation of forest stock is crucial in forest resource management and is one of the important monitoring indicators.In this study,Zhanyi District,Qujing City,Yunnan Province,was taken as the study area,and Pinus yunnanensis was used as the study object to construct a stochastic forest accumulation estimation model based on the spectral features of Landsat9 remote sensing images.The random forest estimation model was opti-mized by grid search algorithm to achieve the regional scale estimation.The results show that the grid search al-gorithm can realize the improvement of the accuracy of the random forest model,and the modeling accuracy of the GS-RF estimation model is:R2=0.67,RMSE=1.11;the validation accuracy is:R2=0.69,RMSE=1.22,and the total accumulation of Pinus yunnanensis in the study area is 3.599 X 106m3.It points out that the overall distri-bution of Pinus yunnanensis is more in the west and east,and less in the center,which may have some correlation with the land use,while the difference of unit accumulation is not obvious,and the overall distribution has some spatial heterogeneity.The GS-RF accumulation estimation model constructed in the study can realize the estima-tion of Pinus yunnanensis accumulation in a large scale,and can provide a method for the estimation of accumula-tion at the regional scale.
NETL
NSTL
万方数据
