The use of UAV remote sensing for studying rockfalls and their impacts on forest

Žabota, Barbara (author), Kobal, Milan (mentor), Berger, Frédéric (comentor)

In this research, we investigated the potential of using UAVs to acquire and produce photogrammetric and LiDAR products that can be used in analyses related to rockfall hazard and protection forest. First, we statistically evaluated the use of direct and indirect georeferencing through bundle block adjustment, a digital surface model, and orthomosaics. The results show that indirect georeferencing is more accurate, but direct georeferencing can be used on a representational scale with the potential to be used with only few ground control points (GCPs). Secondly, we compared field measurements of the dimensions and volumes of rock deposits with UAV photogrammetric and LiDAR point cloud measurements. The measurement results were further used as input parameters in a 3D rockfall model to test whether there are differences between the methods that could influence the planning of protection measures. There were no statistically significant differences between the measurement methods between the rock dimensions and volumes as well as for the modeling results of propagation probability and maximum rebound heights. However, large differences exist for modelled maximum kinetic energies. Finally, we used multiband UAV imagery and used it to calculate vegetation indices (VIs) to investigate whether it is possible to identify trees injured due to rockfall impacts, and differentiate between injury groups. The conclusions show that the identification of trees injured by rockfall (both conifer and deciduous tree species) was possible at study sites where rockfall injuries were at least one year old. It was not possible to distinguish between the injury groups in the injured trees, but there was a strong correlation between the VIs and the extent of the injury according to the basal area of the trees.

• Language: English
• Keywords: rockfalls, unmanned aerial vehicles, remote sensing, photogrammetry, LiDAR, protection forests
• Work type: Doctoral dissertation
• Typology: 2.08 - Doctoral Dissertation
• Organization: BF - Biotechnical Faculty
• Publisher: [B. Žabota]
• Year: 2023
• PID: 20.500.12556/RUL-152017
• UDC: 630*3+630*5:630*4(043.3)=111
• COBISS.SI-ID: 170513155
• Publication date in RUL: 28.10.2023

Secondary language

• Language: Slovenian
• Title: Uporaba daljinskega zaznavanja z brezpilotnimi letalniki za proučevanje skalnih podorov in njihovih vplivov na gozd

Abstract

V okviru raziskave smo preučevali možnosti uporabe daljinskega zaznavanja z brezpilotnimi letalniki za pridobivanje in izdelavo fotogrametričnih in LiDAR-izdelkov, ki se lahko uporabljajo pri analizah skalnih podorov in varovalnega gozda. Z namenom preverjanja prve raziskovalne hipoteze smo statistično ovrednotili uporabo neposrednega in posrednega georeferenciranja izravnave blokov fotografij, digitalnega modela površja in ortomozaikov. Rezultati kažejo, da s posrednim georeferenciranjem dosegamo točnejše fotogrametrične izdelke. Vendar pa bi se lahko neposredno georeferenciranje uporabljalo na reprezentativni ravni in v kombinaciji z manjšim številom talnih kontrolnih točk. V primeru druge raziskovalne hipoteze smo primerjali terenske meritve dimenzij in volumnov skalnih blokov z meritvami v fotogrametričnem in lidarskem oblaku točk, da bi preučili ali obstajajo razlike v modeliranju območij skalnih podorov. Med merilnimi metodami ni bilo statistično značilnih razlik pri dimenzijah in volumnih skalnih gmot, kot tudi ne pri modeliranju maksimalnega dosega skalnih blokov in maksimalne višine odboja skalnih gmot. Razlike pa so prisotne v primeru modelirane maksimalne kinetične energije. V okviru tretje raziskovalne hipoteze smo na podlagi multispektralnih posnetkov izračunali vegetacijske indekse in jih uporabili za prepoznavanje dreves, poškodovanih zaradi skalnih podorov. Ugotovili smo, da je poškodovana drevesa mogoče prepoznati tam, kjer so poškodbe stare vsaj eno leto. Razlikovanje med različnimi skupinami poškodb ni bilo mogoče; ugotovljena pa je bila visoka korelacija med vrednostmi vegetacijskih indeksov in velikostjo poškodbe glede na temeljnico drevesa.

• Keywords: skalni podori, brezpilotni letalniki, daljinsko zaznavanje, fotogrametrija, LiDAR, varovalni gozdovi