The determination of some stand parameters using SfM-based spatial 3D point cloud in forestry studies: an analysis of data production in pure coniferous young forest stands

Benefiting from current unmanned air vehicle (UAV) and remote sensing techniques, the present study aims to estimate tree count (TC), tree height (TH), and tree crown cover area (TCCA) in a young Calabrian pine stand via canopy height model (CHM). Overlay images obtained using Quadcopter were used to generate two spatial three-dimensional (3D) cloud points in two different qualities. Point clouds were processed using R program in order to produce tree data using CHM. The sensitivity of CHM-based tree data was revealed using 318 tree measurements in 32 different sampling units. Estimation and measurement values were classified based on their structure from motion (SfM) quality and cover classes, and the statistical relationships among them were analyzed. Without any classification, R-2 was calculated for TC, THMean, and TCCA(Total) estimations and field measurements. R-2 values were calculated as 0.865, 0.778, and 0.869, respectively, for SfM(Highest) CHM, while they were calculated as 0.863, 0.736, and 0.843, respectively, for SfM(Medium) CHM. In addition, sensitivity and performance ranking in different groups were determined based on root mean square error (RMSE) and mean absolute percentage error (MAPE) values. A significant difference was observed among groups in terms of quality and cover for TH, while no significant differences were observed for TCCA. Therefore, it is possible to estimate the properties of SfM CHM-based young coniferous stand. It was understood that tree density, crown shape, and branching influenced the accuracy of the present study. The developed UAV (Drone)-SfM is a promising technique for further small-scale forestry studies.