Fast and stable direct relative orientation of UAV-based stereo pair

International Journal of Advances in Intelligent Informatics

View Publication Info
Field Value
Title Fast and stable direct relative orientation of UAV-based stereo pair
Creator Tjahjadi, Martinus Edwin
Agustina, Fransisca Dwi
Subject Relative Orientation; Closed-Form Solution; Stereoscopic Processing
Description Coplanarity-based relative orientation (RO) is one of the most crucial processes to obtain reliable 3D model and point clouds in Computer Vision and Photogrammetry community. Whilst a classical and rigorous procedure requires very close approximate values of five independent parameters, a direct method needs additional constraints to solve the parameters. This paper proposes a new approach that facilitates a very fast but stable and accurate solution from five point correspondences between two overlapping aerial images taken form unmanned aerial vehicle (UAV) flight. Furthermore, if 3D coordinates of perspective centers are available form geotagged images, rotational elements of the RO parameters can be quickly solved using three correspondences only. So it is very reliable for a provision of closed-form solutions for the rigorous methods. Our formulation regards Thompson’s parameterizations of Euler angles in composing a coplanarity condition. Nonlinear terms are subsequently added into a stereo parallax within a constant term under a linear least squares criteria. This strategy is considered new as compared with the known literatures since the proposed approach can find optimal solution. Results from real datasets confirm that our method produces a fast, stable and reliable linear solution by using at least five correspondences or even only three conjugate points of geotagged image pairs.
Publisher Universitas Ahmad Dahlan
National Institute of technology (ITN) Malang
Date 2019-03-23
Type info:eu-repo/semantics/article

Format application/pdf
Source International Journal of Advances in Intelligent Informatics; Vol 5, No 1 (2019): March 2019; 24-39
Language eng

Contact Us

The PKP Index is an initiative of the Public Knowledge Project.

For PKP Publishing Services please use the PKP|PS contact form.

For support with PKP software we encourage users to consult our wiki for documentation and search our support forums.

For any other correspondence feel free to contact us using the PKP contact form.

Find Us


Copyright © 2015-2018 Simon Fraser University Library