Stručni rad

https://doi.org/10.19279/TVZ.PD.2024-12-2-07

COMPARISON OF PHOTOGRAMMETRY METHOD AND GAUSSIAN SPLATTING IN 3D MODEL RECONSTRUCTION

Emanuel Dubovečak ; Tehničko veleučilište u Zagrebu, Vrbik 8, 10000 Zagreb, Hrvatska *
Ivan Rajković orcid.org/0000-0002-3340-3418 ; Tehničko veleučilište u Zagrebu, Vrbik 8, 10000 Zagreb, Hrvatska *
Dinka Radonić ; Tehničko veleučilište u Zagrebu, Vrbik 8, 10000 Zagreb, Hrvatska *

* Dopisni autor.

Sažetak

This paper explores the ways in which technological progress in aerial photography and computer vision affect the reconstruction of detailed 3D models using aerial imagery. Technological progress in the field of 3D model reconstruction specifically involves photogrammetry and Gaussian splatting methods. Historically, both methods have undergone significant development. Photogrammetry, known as a pioneering method in 3D model creation, has advanced through digital photography and image processing, employing modern algorithms for model reconstruction. On the other hand, Gaussian splatting, although less familiar, has developed alongside technological advancements in algorithm optimization, opening up new possibilities for visualizing 3D scenes and models. The paper analyses the differences between these two methods and their application in 3D model reconstruction. It considers existing theoretical knowledge and the practical process of reconstructing 3D models using these methods. Ultimately, this study aims to enhance understanding of these methods and technological progress, significantly facilitating the selection of the most appropriate method for visualizing specific 3D scenes and models.

Ključne riječi

technology; digital photography; 3D model; photogrammetry, Gaussian splatting

Hrčak ID:

329249

URI

https://hrcak.srce.hr/329249

Datum izdavanja:

20.11.2024.

Podaci na drugim jezicima: hrvatski

Posjeta: 0 *