What question did this study set out to answer?

This research aims to evaluate Neural Radiance Fields for 3D reconstruction compared to standard SfM-MVS techniques.

January 14, 2026GeomaticsOpen Access

Evaluating Neural Radiance Fields for Image-Based 3D Reconstruction: A Comparative Study with SfM-MVS

View Full Paper

Ask AI

Bookmark

Authors

Comparative analysis reveals that Neural Radiance Fields show promising results in 3D modelling, suggesting potential for diverse applications.

This research aims to evaluate Neural Radiance Fields for 3D reconstruction compared to standard SfM-MVS techniques.
Evaluated performance of Neural Radiance Fields via Nerfstudio and SfM-MVS using Agisoft Metashape Professional.
Considered various datasets with differing object scales and capture methods, including drone imagery.
Conducted assessments of accuracy, completeness, planarity, and density of reconstructed point clouds.
Neural Radiance Fields produced reconstructions comparable to conventional methods with accurate camera poses.
Promising spatial consistency was observed despite the initial lack of metric accuracy.
Influences of shadows and surface flatness on reconstruction fidelity were analyzed.

Giaquinto et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark

Comparative analysis reveals that Neural Radiance Fields show promising results in 3D modelling, suggesting potential for diverse applications.