Individual Carpinus betulus and Acer velutinum tree species delineation by geometrical and statistical characteristics derived from airborne LiDAR data

In this research, the possibility of application of LiDAR-derived characteristics has been studied for discrimination amongst common hornbeam (Carpinus betulus) and Persian maple (Acer velutinum) trees. LiDAR data of individual sample trees were separated in laser point clouds using their measured center coordinates and crown diameter in the field. In district 1 of Shast-Kolate Education and Research Forest in Gorgan, 80 individual A. velutinum and C. betulus tree samples were selected. The trees were either located in dominant storey or were not overlaid by adjacent tree crowns. Tree heights were measured using Vertex 1V GPS device. Crown diameter was measured in four cardinal directions using Leica Disto lasermeter. Center coordinates of the sample trees were determined using both DGPS and distance/azimuth measurement by Total Station device. Different geometrical and statistical metrics of sample trees were extracted from LiDAR data. The results of discriminant analysis suggested the height standard deviation of laser points over 80% of tree height and crown slope as the selected metrics to differentiate the tree species (accuracy=80.3%). The mean of those two metrics showed larger values for hornbeam than maple. The selected LiDAR metrics were ascribed to represent the shape and height variation of returned crown laser points. It is therefore concluded that geometrical LiDAR-derived attributes could successfully helo by differentiating between the two tested tree species.