A Comparison of Multi-View Stereo Methods for Photogrammetric 3D Reconstruction: From Traditional to Learning-Based Approaches
Y. Li 1, G. Vosselman 1, F. Nex 1 · 2026
Read A Comparison of Multi-View Stereo Methods for Photogrammetric 3D Reconstruction: From Traditional to Learning-Based Approaches by Y. Li 1, G. Vosselman 1, F. Nex 1 (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-10246.txt.
Description
Li, Vosselman, and Nex compare traditional and learning-based multi-view stereo for photogrammetric reconstruction. Traditional COLMAP-style reconstruction gives high accuracy but runs slowly. Learning methods are geometry-guided or end-to-end, tested on two aerial scenes. COLMAP stays reliable and consistent but slow, while learning methods match features better and are more robust where traditional registration fails. Geometry-guided methods need preparation and COLMAP priors, while end-to-end DUSt3R and VGGT match COLMAP accuracy and coverage faster, though with larger residuals in hard scenes.
Catalog record
| CREATOR | Y. Li 1, G. Vosselman 1, F. Nex 1 |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 32592 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.10246