Estimation of qualitative and quantitative characteristics of Pistacia atlantica Desf. and Amygdalus spp. in of UAV point clouds

Estimation of qualitative and quantitative characteristics of plants on UAV images is considered a challenge in remote sensing. Therefore, this study aimed to present a method to estimate crown area, height, and species in a mixed Pistacia-Amygdalus stand in UAV-derived point clouds. To this aim, 100 Pistacia atlantica Desf. trees and 100 Amygdalus spp < em>. shrubs were randomly selected. Point cloud was obtained by UAV-derived imagery with 50 points per m2 in a 64-ha study area in Baneh Research Forest, Fars province. The quantitative characteristics were then estimated on the point cloud. Additionally, species type was classified using random forest and 37 quantitative attributes measured on point cloud, canopy height model, and orthomosaic. Crown area and height of Pistacia (R2= 0.91 and 0.83, PRMSE=4.7% and 3.2%, respectively) and Amygdalus (R2= 0.89 and 0.47, PRMSE=22.1% and 21.5%, respectively) were also estimated. By application of quantitative attributes and random forest, species type was classified with an accuracy of 0.92 and κ of 0.98. All in all, results indicated that UAV point clouds can be efficiently applied to estimate a set of qualitative and quantitative attributes of Pistacia and Amygdalus within the study area. However, inaccurate and imprecise results were observed for estimated heights of Amygdalus.