Calibration and Accuracy Assessment in a Direct Georeferencing System in UAS Photogrammetry

Petr Gabrlik, Anders La Cour-Harbo, Petra Kalvodova, Ludek Zalud, Premysl Janata

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review

16 Citationer (Scopus)
346 Downloads (Pure)

Resumé

Unmanned aerial systems (UASs) have been already proven to be useful in fields and disciplines such as agriculture, forestry, or environmental mapping, and they have also found application during natural and nuclear disasters. In many cases, the environment is inaccessible or dangerous for a human being, meaning that the widely used technique of aerial imagery georeferencing via ground control points cannot be employed. The present article introduces a custom-built multi-sensor system for direct georeferencing, a concept that enables georeferencing to be performed without an access to the mapping area and ensures centimetre-level object accuracy. The proposed system comprises leading navigation system technologies in the weight category of micro and light UASs. A highly accurate Global Navigation Satellite System receiver integrating the real-time kinematic technology supports an inertial navigation system, where data from various sensors are fused. Special attention is paid to the time synchronization of all sensors, and a method for the field calibration of the system is designed. The multi-sensor system is completely independent of the used UASs. The authors also discuss the verification of the proposed system’s performance on a real mission. To make the results credible, a high number of test points are used, with both direct and indirect georeferencing techniques subjected to comparison, together with different calibration methods. The achieved spatial object accuracy (about 4 cm root mean square error (RMSE)) is sufficient for most applications.
OriginalsprogEngelsk
TidsskriftInternational Journal of Remote Sensing
Vol/bind39
Udgave nummer15-16
Sider (fra-til)4931-4959
Antal sider29
ISSN0143-1161
DOI
StatusUdgivet - 10 feb. 2018

Fingerprint

accuracy assessment
photogrammetry
sensor
calibration
navigation
GNSS
ground control
disaster
forestry
imagery
kinematics
agriculture
method

Citer dette

Gabrlik, Petr ; La Cour-Harbo, Anders ; Kalvodova, Petra ; Zalud, Ludek ; Janata, Premysl. / Calibration and Accuracy Assessment in a Direct Georeferencing System in UAS Photogrammetry. I: International Journal of Remote Sensing. 2018 ; Bind 39, Nr. 15-16. s. 4931-4959.
@article{bd8291220245420a87268021da12b207,
title = "Calibration and Accuracy Assessment in a Direct Georeferencing System in UAS Photogrammetry",
abstract = "Unmanned aerial systems (UASs) have been already proven to be useful in fields and disciplines such as agriculture, forestry, or environmental mapping, and they have also found application during natural and nuclear disasters. In many cases, the environment is inaccessible or dangerous for a human being, meaning that the widely used technique of aerial imagery georeferencing via ground control points cannot be employed. The present article introduces a custom-built multi-sensor system for direct georeferencing, a concept that enables georeferencing to be performed without an access to the mapping area and ensures centimetre-level object accuracy. The proposed system comprises leading navigation system technologies in the weight category of micro and light UASs. A highly accurate Global Navigation Satellite System receiver integrating the real-time kinematic technology supports an inertial navigation system, where data from various sensors are fused. Special attention is paid to the time synchronization of all sensors, and a method for the field calibration of the system is designed. The multi-sensor system is completely independent of the used UASs. The authors also discuss the verification of the proposed system’s performance on a real mission. To make the results credible, a high number of test points are used, with both direct and indirect georeferencing techniques subjected to comparison, together with different calibration methods. The achieved spatial object accuracy (about 4 cm root mean square error (RMSE)) is sufficient for most applications.",
author = "Petr Gabrlik and {La Cour-Harbo}, Anders and Petra Kalvodova and Ludek Zalud and Premysl Janata",
year = "2018",
month = "2",
day = "10",
doi = "10.1080/01431161.2018.1434331",
language = "English",
volume = "39",
pages = "4931--4959",
journal = "International Journal of Remote Sensing",
issn = "0143-1161",
publisher = "Taylor & Francis",
number = "15-16",

}

Calibration and Accuracy Assessment in a Direct Georeferencing System in UAS Photogrammetry. / Gabrlik, Petr; La Cour-Harbo, Anders; Kalvodova, Petra; Zalud, Ludek; Janata, Premysl.

I: International Journal of Remote Sensing, Bind 39, Nr. 15-16, 10.02.2018, s. 4931-4959.

Publikation: Bidrag til tidsskriftTidsskriftartikelForskningpeer review

TY - JOUR

T1 - Calibration and Accuracy Assessment in a Direct Georeferencing System in UAS Photogrammetry

AU - Gabrlik, Petr

AU - La Cour-Harbo, Anders

AU - Kalvodova, Petra

AU - Zalud, Ludek

AU - Janata, Premysl

PY - 2018/2/10

Y1 - 2018/2/10

N2 - Unmanned aerial systems (UASs) have been already proven to be useful in fields and disciplines such as agriculture, forestry, or environmental mapping, and they have also found application during natural and nuclear disasters. In many cases, the environment is inaccessible or dangerous for a human being, meaning that the widely used technique of aerial imagery georeferencing via ground control points cannot be employed. The present article introduces a custom-built multi-sensor system for direct georeferencing, a concept that enables georeferencing to be performed without an access to the mapping area and ensures centimetre-level object accuracy. The proposed system comprises leading navigation system technologies in the weight category of micro and light UASs. A highly accurate Global Navigation Satellite System receiver integrating the real-time kinematic technology supports an inertial navigation system, where data from various sensors are fused. Special attention is paid to the time synchronization of all sensors, and a method for the field calibration of the system is designed. The multi-sensor system is completely independent of the used UASs. The authors also discuss the verification of the proposed system’s performance on a real mission. To make the results credible, a high number of test points are used, with both direct and indirect georeferencing techniques subjected to comparison, together with different calibration methods. The achieved spatial object accuracy (about 4 cm root mean square error (RMSE)) is sufficient for most applications.

AB - Unmanned aerial systems (UASs) have been already proven to be useful in fields and disciplines such as agriculture, forestry, or environmental mapping, and they have also found application during natural and nuclear disasters. In many cases, the environment is inaccessible or dangerous for a human being, meaning that the widely used technique of aerial imagery georeferencing via ground control points cannot be employed. The present article introduces a custom-built multi-sensor system for direct georeferencing, a concept that enables georeferencing to be performed without an access to the mapping area and ensures centimetre-level object accuracy. The proposed system comprises leading navigation system technologies in the weight category of micro and light UASs. A highly accurate Global Navigation Satellite System receiver integrating the real-time kinematic technology supports an inertial navigation system, where data from various sensors are fused. Special attention is paid to the time synchronization of all sensors, and a method for the field calibration of the system is designed. The multi-sensor system is completely independent of the used UASs. The authors also discuss the verification of the proposed system’s performance on a real mission. To make the results credible, a high number of test points are used, with both direct and indirect georeferencing techniques subjected to comparison, together with different calibration methods. The achieved spatial object accuracy (about 4 cm root mean square error (RMSE)) is sufficient for most applications.

U2 - 10.1080/01431161.2018.1434331

DO - 10.1080/01431161.2018.1434331

M3 - Journal article

VL - 39

SP - 4931

EP - 4959

JO - International Journal of Remote Sensing

JF - International Journal of Remote Sensing

SN - 0143-1161

IS - 15-16

ER -