Study On The Remote Sensing Feature Selection Method For Forest Biomass Estimation Based On RF-RFE

Forest biomass estimation is the key to maintain the global carbon balance and protect the environment. Some characteristic parameters associated with the forest biomass can be extracted from the remote sensing image, including the single-band information, vegetation indexes, texture features and terrain factors, which have been used to estimate the forest biomass accurately. Feature selection can effectively reduce the redundant remote sensing of forest biomass in the process of the characteristics of redundancy, Which can reduce the error of the transfer to a certain extent and provide some reference for remote sensing. However, the quantity of characteristic parameters is huge, and too many parameters have a negative influence on the precision of prediction. A novel parameter-selection method that combines the Random Forest and Recursive Feature Elimination(RF-RFE) is proposed to reduce the number of biomass parameters, and enhance the estimation accuracy of forest biomass. This study adopted the ZY-3Satellite Image and the plot data of forest inventory in 2012, and used RF-RFE to select the parameters and estimate the forest biomass in Jiliuhe forest farm in Greater Khingan Mountain. The main research contents and results are as follows:(1) Plot biomass calculation in Jiliuhe forest farm in Greater Khingan MountainBy December 2012 Daxinganling three kinds of investigation data of forest, tree species composition, hectare, small area data, through consulting relevant literature on the use of biomass conversion factor function calculation. Finally, it obtain the true value of the biomass of 87 samples.(2) Forest type recognition based on texture featuresIn this paper, we use the method of supervised classification of support vector machine based on texture feature to accomplish the recognition of forest. The proportion of larch, birch and mixed forests was 9.78%, 28.66% and 32.16% respectively.(3) Remote Sensing Feature Selection Method for Forest Biomass Estimation Based on RF-RFEThe data of 87 samples of three kinds of survey data in Greater Khingan Range and the data of three kinds of remote sensing image that were processed were analyzed. RF-RFE algorithm was used to select the 49 characteristics of the 87 sample plots Jiliuhe forest farm. The final selection results are selected according to the size of R2 and RMSE. Unclassified, feature selection feature number is 10. After classification, sample is the result of larch, birch and mixed forest were 6, 5, 7.(4) Model comparative analysisRegression model and support vector machine were used to select the characteristics of forest biomass remote sensing in the study area. Then the results are compared with the results of the algorithm proposed in this paper. The results show that the algorithm proposed in this paper is more effective and universal.