Estimation of some stand parameters using aerial photographs and sentinel-2 images

In forestry, aerial photographs can be used to generate 2D orthophotos for area inventory works and/or to create 3D point clouds, especially for tree height estimations. However, no comprehensive study has been conducted on the basis of planning units in our country. The present study aims to minimize forest field inventory by remotely sensed data sources. Based on the ground measurement from 170 temporary sample plots (119 plots model, 51 plot test), the number of stem, stand volume, crown closure, mean tree height and mean DBH were estimated using the HGM's aerial photographs and Sentinel-2 satellite image reflectance values. The study was conducted in Antalya Regional Directorate of Forestry, Alanya State Forest Enterprise, Alara Planning Unit, which was dominated by Turkish Red Pine (Pinus brutia) tree species. The digital surface model (DSM) and digital terrain model (DTM) were produced using point clouds derived from aerial photographs, and then the canopy height model (CHM) was created by extracting DTM from DSM. Afterwards, correlation and regression analyses were performed between the stand parameters and CHM, and Sentinel-2 data. The results showed that the correlation coefficient (Pearson's r) for the mean tree height component was 0,76 based on CHM and ground measurement. According to multi linear regression models, the adjusted coefficients of determination (R2) of the mean tree height, stand volume, number of stem and mean DBH were found to be 0,61; 0,45; 0,21 and 0,43 with RMSEs of 2,4 m; 57,4 m3/ha; 161 stems/ha and 6,1 cm, respectively.