Application of Gaofen-2 Remote Sensing Imagery in Bamboo Forest Extraction:a Case Study of Huangfengqiao Forest Farm
WU Kang
GU Jun
LIN Hui
Abstract：This study used the High-Resolution II remote sensing imagery,combined with spectral features,texture features,and vegetation indices,to precisely extract and analyze the distribution of bamboo forests in Huangfengqiao Forest Farm in Hunan Province using machine learning methods.First,by conducting in-depth analysis of the spectral reflectance differences between bamboo forests and other main tree species at different times,it was determined that spring was the most significant period for the spectral features of bamboo forests,and the spring High-Resolution II imagery was selected as the research data source.The second-order probability statistical co-occurrence matrix method was used to extract multiple texture features,followed by the Pearson correlation analysis and the random forest recursive feature elimination(RF-RFE)method to select the 13 most relevant features.Finally,three classification algorithms(RF,SVM,and KNN)were used to generate a bamboo forest distribution map,the experimental results showed that the random forest algorithm performed best,with an overall accuracy of 94.83%.Although this study relied solely on spring single-time-phase remote sensing data,future studies can incorporate large-scale,multi-temporal imagery to further improve the accuracy and timeliness of bamboo forest distribution extraction.
Keywords：moso bamboo forest extractionvegetation indexBest of the timeRFSVMKNN
Publication Date：2025-04-28
Online Publishing Date：2026-05-22(First online date of this platform, not the publication date of the document)
Pages：7( 43-49 )