Estimation and Analysis of Forest Resource Planning and Design Survey Volume Supported by Multi-Source Data
This study takes a coniferous broad leaved mixed forest in XX city,Guangdong province as the research area based on Sentinel-1 and Sentinel-2 remote sensing data.Combined with ground measurement data,a volume estimation model is constructed.The research results indicate that using Sentinel-1 data+Sentinel-2 data as the data source,using the distance correlation coefficient method to screen features,and using the random forest mod-el to construct the volume estimation model has the highest accuracy,with a relative root mean square error(RRMSE)of 0.2217 and a decision coefficient R2 0.61.The model fitting effect is good;The predicted volume is concentrated between 300 m3/hm2 and 400 m3/hm2.Paired t-tests were conducted with actual values,and the re-sults showed that there was no significant difference between the predicted and actual values(t=0.001,P=0.999＞0.01),further proving the reliability of the model.
Sentinel-1 dataSentinel-2 dataforest resource planning and design surveyVolume estimation
NETL
NSTL
万方数据
