Fotogrametrično merjenje višine dreves z daljinsko vodenim letalnikom - možnosti in omejitve

Perme, Gašper (author), Kobal, Milan (mentor)

V diplomski nalogi smo primerjali točnost izmere fotogrametrične višine 36 dreves, ki so rasla na samem in v ravnini v primerjavi s tahimetrično izmerjenimi višinami, ki so prestavljale referenčne podatke. Podatke za izdelavo oblakov točk smo pridobili iz fotografij, zajetih z daljinsko vodenim letalnikom. Rezultati analize so pokazali, da se višina dreves, izračunana iz digitalnega modela krošenj (DMK), ki je nastal na podlagi digitalnega modela površja (DMP) in digitalnega modela reliefa (DMR), statistično značilno razlikuje (parni t-test; p < 0,05) od referenčnih vrednosti, ne glede na vir DMR. Pri metodi, kjer sta bila tako DMP kot DMR izdelana iz fotogrametričnega oblaka točk, je bila povprečna razlika -0,38 m (95 % interval zaupanja med -0,48 m in -0,29 m; RMSE = 0,47 m). Tudi v primeru kombinirane metode, kjer je bil DMR pridobljen iz lidarskih podatkov, je bila razlika statistično značilna (parni t-test; p < 0,05), vendar manjša: -0,11 m (95 % interval zaupanja med -0,22 m in -0,01 m; RMSE = 0,32 m). Primerjava višin dreves, izračunanih iz fotogrametričnega DMK in z obema metodama, je pokazala statistično značilne razlike med njimi (p < 0,05), pri čemer je bila povprečna razlika 0,27 m (95 % interval zaupanja med 0,20 m in 0,34 m). Kljub temu pa se točnost izračuna višine glede na skupino dreves (iglavci vs. listavci) ni statistično značilno razlikovala za nobeno od metod (p > 0,05), kar kaže, da je vpliv drevesne vrste na izračun višine drevesa zanemarljiv v obeh primerih. Zaradi nizkih vrednosti RMSE lahko zaključimo, da je fotogrametrija uporabna za ocenjevanje višine dreves, tako iglavcev kot listavcev, vendar je pri interpretaciji potrebno upoštevati izbrano fotogrametrično metodo merjenja.

• Language: Slovenian
• Keywords: višina dreves, fotogrametrija, LiDAR, daljinsko vodeni letalniki
• Work type: Bachelor thesis/paper
• Typology: 2.11 - Undergraduate Thesis
• Organization: BF - Biotechnical Faculty
• Publisher: [G. Perme]
• Year: 2025
• PID: 20.500.12556/RUL-170575
• UDC: 630*58(043.2)=163.6
• COBISS.SI-ID: 242409731
• Publication date in RUL: 10.07.2025

Secondary language

• Language: English
• Title: Photogrametric tree height measurement using a remotely piloted aircraft system - possibilities and limitations

Abstract

In this thesis, we evaluated the accuracy of photogrammetric tree height measurements for 36 lone trees growing on flat terrain by comparing them with reference heights determined by total station surveying. The data for creating the point cloud was obtained from photos taken with a remotely piloted aircraft system (RPAS). The analysis showed that the tree heights calculated from the canopy height model (CHM) obtained by subtracting the digital terrain model (DTM) from the digital surface model (DSM) deviated significantly from the reference values, regardless of the DTM source. When both the DSM and the DTM were generated from a photogrammetric point cloud, the average difference to the reference was 0.38 m (95% confidence interval: -0.48 m to -0.29 m; RMSE = 0.47 m). Using a combined method, where the DTM was generated from a LiDAR point cloud and the DSM from photogrammetry, the average difference was smaller, but still statistically significant: -0.11 m (95% CI: -0.22 m to -0.01 m; RMSE = 0.32 m). A comparison of tree heights calculated from photogrammetric CHM and using both methods showed statistically significant differences between them (p < 0.05), with an average difference of 0.27 m (95% CI: 0.20 m to 0.34 m). On the other hand, the accuracy of the height estimation did not differ significantly between coniferous and broadleaf trees for either method (p > 0.05), which indicates that the tree type had no significant influence on the measurement accuracy. Based on the relatively low root mean square error (RMSE), we conclude that photogrammetry is a suitable method for estimating tree height, for both conifers and deciduous trees, although the choice of photogrammetric processing method must be carefully considered when interpreting the results.

• Keywords: tree height, photogrammetry, LiDAR, remotely piloted aircraft system