With the development of surface imaging technology, data acquisition (photogrammetry, airborne laser scanning) by unmanned aerial vehicle (UAV) is becoming increasingly affordable. In this thesis, we focus on the use of photogrammetry to calculate growing stocks using several different regression models, where we consider as explanatory variables the measurements of trees in the field and the data that can be obtained from the Canopy Height Model (CHM). Field measurements of trees (31 sample plots, 454 trees) and photographs (542) were taken with a DJI M300 RTK unmanned aerial aircraft in the Gorjanci Forest Estate Gorjanci - Hren, section 105A. The results show that the model prediction of the growing stock is more accurate if the model uses field measurement data based on the 700 m2 plots (basal area) in addition to the CHM stand canopy height data, i. e. RMSE = 22.0 m3/ha or 4.9%. If the model uses, in addition to the CHM data on stand canopy height, the data on stand basal area calculated using data gathered with Bitterlich’s angle count sampling method, the result amounts to RMSE = 74.0 m3/ha, or 16.4%. If the model prediction of the growing stock takes into account only the canopy height model (CHM) (average stand canopy height, canopy closure), the RMSE is 89.4 m3/ha, or 19.9%.
|
