Mapping tree species in a boreal forest area using RapidEye and LiDAR data

Tree species composition is an indicator of forest type. It is also a required attribute in forest inventory, biomass and stand volume estimation. Accurate mapping tree species is essential for forest management purposes. In this paper the performances of LiDAR, RapidEye data, and their combination on tree species classification were investigated in a boreal forest. Both Random forest (RF) and support vector machine (SVM) classification methods were performed. Results indicated that combined LiDAR and RapidEye data improved the classification accuracy significantly, compare to using each type of data separately. The RF classifier outperformed SVM for tree species classification. Six variables that contributed most to classification accuracy were digital elevation model, slope, canopy height, red-edge NDVI, and red-edge and Near infrared bands of RapidEye data.