Metsaökosüsteemide looduslikkuse hindamine kaugseiremeetoditega
Laen...
Kuupäev
2023
Kättesaadavus
06.09.2023
Autorid
Ajakirja pealkiri
Ajakirja ISSN
Köite pealkiri
Kirjastaja
Eesti Maaülikool
Abstrakt
Metsade bioloogilise mitmekesisuse vähenemise peamiseks põhjuseks on metsade
intensiivne majandamine ja seeläbi metsaökosüsteemis elupaikade kadumine. Selle
ärahoidmiseks tuleb tõsta sellealast teadlikkust ning leida viise, kuidas lihtsamini ja
kiiremini koguda infot metsa looduslikkuse kohta. Käesoleva bakalaureusetöö eesmargiks
on uurida, kas LiDARi ning fotogramm-meetrilist kaugseirevahendeid kasutades on
võimalik tuvastada metsade looduslikkuse seisundit. Töös võrreldi omavahel MA-ALS
(Manned Aircraft Airborne Laser Scanning) ja UAV-ALS (Unmanned Aerial Vehicle
ALS) mõõdetud andmeid ning lisaks uuriti ka UAV DAP (Digital Aerial Phogrammetry)
meetodil mõõdetud andmeid. Mõõtmised teostati 30 eriloodusväärtuslikes puistutes, mis
asusid Jõgeva, Tartu ja Põlva maakonnas. Saadud andmetest koostati punktipilved, millest
arvutati välja statistilised kõrgusmeetrikud ning tehti nendega regressioonianalüüse.
DAP meetodi andmete töötluse käigus selgus, et tulemust ei ole võimalik saavutada, kuna
pildistamise ajal kõikusid puud liiga suures ulatuses ning puude detailne struktuur
võimendas piltide ühildamise ebakõla veelgi. Probleemi saaks lahendada, kui teiste
autorite eeskujul võtta kasutusse silmapaistvad maapealsed tähised.
LiDAR andmetest tuli välja, et UAV-ALS andmetest saab koostada täpsemaid mudeleid
kui MA-ALS, mida toetavad ka teiste autorite tööd. LiDAR andmetest on võimalik
koostada prognoosimudeleid loodusväärtuse hindamiseks, kuid eri puistute põhiselt tuleb
jälgida, milliseid mudeleid kasutada, et andmeid õigesti mõista. Puuliigi ning metsatüübi
spetsiifilised mudelid näitavad potentsiaali olla kasulikud, kuid selleks peab neid suurema
valimi sees edaspidi analüüsima.
The main cause of the decline in forest biodiversity is the intensive management of forests, resulting in the loss of habitat in forest ecosystems. To prevent this, it is necessary to raise awareness and find ways to gather information about forest naturalness more easily and quickly. The aim of this bachelor's thesis is to investigate whether LiDAR and photogrammetric remote sensing methods can be used to detect the naturalness status of forests. The study compared Manned Aircraft airborne Laser Scanning (MA-ALS) and UAV-based (Unmanned Aerial Vehicle) LiDAR data, and also examined data obtained through UAV digital aerial photogrammetry (DAP). Measurements were carried out in 30 sites of different natural value forests located in Jõgeva, Tartu, and Põlva counties. Point clouds were generated from the collected data, and statistical height metrics were computed and subjected to regression analysis. During the data processing of the DAP method, it was found that achieving accurate results was not possible due to significant variations in tree positions by the wind at the time of imaging, and the detailed tree structure further exacerbated the inconsistency in image alignment. The problem could be addressed by adopting prominent ground-based markers, following the example of other authors. Results from LiDAR data revealed that UAV-based LiDAR scanning can generate more accurate models than MA-ALS, which is supported by the work of other authors as well. It is possible to create predictive models for assessing natural values based on LiDAR data, but it is essential to consider using specific models for different forest types and tree species to correctly interpret the data. These specific models for tree species and forest types show potential usefulness, but further analysis within a larger sample size is needed to explore their application.
The main cause of the decline in forest biodiversity is the intensive management of forests, resulting in the loss of habitat in forest ecosystems. To prevent this, it is necessary to raise awareness and find ways to gather information about forest naturalness more easily and quickly. The aim of this bachelor's thesis is to investigate whether LiDAR and photogrammetric remote sensing methods can be used to detect the naturalness status of forests. The study compared Manned Aircraft airborne Laser Scanning (MA-ALS) and UAV-based (Unmanned Aerial Vehicle) LiDAR data, and also examined data obtained through UAV digital aerial photogrammetry (DAP). Measurements were carried out in 30 sites of different natural value forests located in Jõgeva, Tartu, and Põlva counties. Point clouds were generated from the collected data, and statistical height metrics were computed and subjected to regression analysis. During the data processing of the DAP method, it was found that achieving accurate results was not possible due to significant variations in tree positions by the wind at the time of imaging, and the detailed tree structure further exacerbated the inconsistency in image alignment. The problem could be addressed by adopting prominent ground-based markers, following the example of other authors. Results from LiDAR data revealed that UAV-based LiDAR scanning can generate more accurate models than MA-ALS, which is supported by the work of other authors as well. It is possible to create predictive models for assessing natural values based on LiDAR data, but it is essential to consider using specific models for different forest types and tree species to correctly interpret the data. These specific models for tree species and forest types show potential usefulness, but further analysis within a larger sample size is needed to explore their application.
Kirjeldus
Bakalaureusetöö
Metsanduse õppekaval
Märksõnad
bakalaureusetööd, lidar, UAV-ALS, DAP, droon, looduslikkus