Mehitamata õhusõidukiga pildistatud aerofotodest valmistatud ortofotomosaiikide ja 3D-punktipilvede täpsust mõjutavad faktorid
Laen...
Kuupäev
2019
Kättesaadav alates
10.09.2019
Autorid
Siim, Karmo
Rahu, Oscar
Ajakirja pealkiri
Ajakirja ISSN
Köite pealkiri
Kirjastaja
Eesti Maaülikool
Abstrakt
Uurimustöö eesmärk oli uurida, kuidas mõjutavad erinevad faktorid ortofotomosaiikide ja
3D-punktipilvede täpsusi, mille andmete kogumiseks on kasutatud mehitamata
õhusõidukit Trimble ZX5. Faktorid, mida uuriti, olid õhusõiduki lennukiirus, mehitamata
õhusõiduki lennukõrgus ja lennumarsruudil pildistatud aerofotode omavaheline kattuvus.
Aeropildistamine toimus Tartu linnas, Eesti Maaülikooli Metsamaja läheduses. Andmete
omavaheliseks võrdlemiseks mõõdeti ortofotomosaiikidelt plaanilised koordinaadid
(X,Y) ja 3D-punktipilvedelt kõrguslikud koordinaadid (Z). Andmete kogumine toimus
2018. aasta sügisel, kuid üks korduslend on tehtud ka 2019. aasta kevadel.
Autorite poolt töös püstitatud hüpoteesideks olid, et kõige täpsemad andmed kogutakse
50 m kõrguselt, piltide kattuvusega 80% × 80% ja lennukiirusega 300 cm/s. Kuna sooviti
teada, millised faktorid annavad fotogramm-meetrilistel töödel kõige paremad tulemused
siis arvutati keskmised ruutvead ja standardhälbed. Selgus, et parimad tulemused tulevad,
kui kasutada aerofotode pildistamiseks 50 m lennukõrguseid (krv X-0,012 m; krv Y-0,012
m; krv Z-0,021 m) kuni 70 m (krv X-0,011 m; krv Y-0,016 m; krv Z-0,020 m),
lennukiirust 400 cm/s (krv X-0,011 m, krv Y-0,012 m; krv Z-0,011 m) kuni 500 cm/s (krv
X-0,010 m; krv Y-0,012 m; krv Z-0,017 m) ja piltide omavahelist kattuvust 80% × 80%
(krv X-0,012 m; krv Y-0,012 m; krv Z-0,021 m) või 70% × 70% (krv X-0,012 m; krv Y0,012 m; krv Z-0,019 m).
Uurimistöö käigus selgus, et mehitamata õhusõidukiga kogutud andmed annavad piisava
täpsuse. Parimate tulemuste parameetrite kasutamine võimaldab valmistada suurema
täpsusega ortofotomosaiike ja 3D-punktipilvesid või 3D-mudeleid.
The goal of this thesis was to determine how different factors affect the accuracy of orthophotomosaics and 3D point clouds, for which data was collected by using a unmanned aerial vehicle (UAV), Trimble ZX5. The following factors were investigated: the flight speed and height of the UAV and the overlapping of the aerial photos during flight. The aerial photography took place in Tartu, near the Institute of Forestry and Rural Engineering of the Estonian University of Life Sciences. For comparison of the data, the coordinates X and Y were measured from the orthophotomosaics and height coordinates Z were measured from the 3D point clouds. The data was collected in autumn 2018, but another flight was conducted in spring 2019. The authors’ hypothesis was that most accurate results would be collected at the height of 50 m, with an 80% × 80% image overlapping and a flight speed of 300 cm/s. To determine which factors provide the best results for photogrammetric projects, the root-mean-square errors (RMSE) and standard deviations were calculated. It was determined that the best conditions for taking aerial photos include a flight height between 50 m (RMSE X-0,012 m; RMSE Y-0,012 m; RMSE Z-0,021 m) to 70 m (RMSE X-0,011 m; RMSE Y-0,016 m; RMSE Z-0,011 m), while flying at a speed between 400 cm/s (RMSE X-0,011 m, RMSE Y-0,012 m; RMSE Z-0,011 m) kuni 500 cm/s (RMSE X-0,010 m; RMSE Y-0,012 m; RMSE Z-0,017 m), using a image overlapping of either 80% × 80% (RMSE X-0,012 m; RMSE Y-0,012 m; RMSE Z-0,021 m) or 70% × 70% (RMSE X-0,012 m; RMSE Y-0,012 m; RMSE Z-0,019 m). The results of the thesis determine that the data collected with the UAV give a sufficient accuracy. Using the parameters with better results enables creating orthophotomosaics and 3D point clouds or 3D models with a higher accuracy.
The goal of this thesis was to determine how different factors affect the accuracy of orthophotomosaics and 3D point clouds, for which data was collected by using a unmanned aerial vehicle (UAV), Trimble ZX5. The following factors were investigated: the flight speed and height of the UAV and the overlapping of the aerial photos during flight. The aerial photography took place in Tartu, near the Institute of Forestry and Rural Engineering of the Estonian University of Life Sciences. For comparison of the data, the coordinates X and Y were measured from the orthophotomosaics and height coordinates Z were measured from the 3D point clouds. The data was collected in autumn 2018, but another flight was conducted in spring 2019. The authors’ hypothesis was that most accurate results would be collected at the height of 50 m, with an 80% × 80% image overlapping and a flight speed of 300 cm/s. To determine which factors provide the best results for photogrammetric projects, the root-mean-square errors (RMSE) and standard deviations were calculated. It was determined that the best conditions for taking aerial photos include a flight height between 50 m (RMSE X-0,012 m; RMSE Y-0,012 m; RMSE Z-0,021 m) to 70 m (RMSE X-0,011 m; RMSE Y-0,016 m; RMSE Z-0,011 m), while flying at a speed between 400 cm/s (RMSE X-0,011 m, RMSE Y-0,012 m; RMSE Z-0,011 m) kuni 500 cm/s (RMSE X-0,010 m; RMSE Y-0,012 m; RMSE Z-0,017 m), using a image overlapping of either 80% × 80% (RMSE X-0,012 m; RMSE Y-0,012 m; RMSE Z-0,021 m) or 70% × 70% (RMSE X-0,012 m; RMSE Y-0,012 m; RMSE Z-0,019 m). The results of the thesis determine that the data collected with the UAV give a sufficient accuracy. Using the parameters with better results enables creating orthophotomosaics and 3D point clouds or 3D models with a higher accuracy.
Kirjeldus
Bakalaureusetöö
Geodeesia ja maakorralduse õppekaval
Märksõnad
bakalaureusetööd, droonid, ortofotomosaiik, fotogramm-meetria